The Politics behind Piketty

by Elizabeth Anderson on December 7, 2015

Thomas Piketty traces widening inequality in rich countries since the early 1970s to increasing shares of income claimed by the top 1%. This trend is decomposed into the increasing share of income accruing to capital ownership, and the increasing share of labor income claimed by corporate executives and financiers. Piketty shows that the increasing share of labor income claimed by the top 1% is neither deserved nor economically useful, in the sense of stimulating better products and services, increasing economic growth, or providing other benefits to the 99%. Because he defines r, the return on capital, as the pure return to passive ownership (excluding returns to capital that could be traced to entrepreneurial activity or business judgment), it is evident that capital’s share of income is also undeserved. But is it economically useful? Piketty misses an opportunity to connect his analysis to a critique of the ideology and associated politics that have driven increasing inequality since the early 1970s. While he rightly claims that the distribution of income and wealth is a deeply political matter, and connects increasing economic inequality to the increasing political clout of the top 1%, he does not identify political decisions, other than cuts in marginal tax rates on top incomes, that lie behind inequality trends. Filling in the ideological and political stories gives us some clues as to policy instruments, other than the tax code, needed to reverse the ominous trends he documents.

On the ideological front, several theories served to rationalized policy shifts in favor of increasing capital shares and top labor incomes. The stagflation of the 1970s was successfully blamed on Keynesian economics, fiscal irresponsibility, a bloated welfare state, militant labor unions, state regulation of the economy, and supposedly incentive-destroying high marginal tax rates on capital incomes and the rich. At the same time, the ideology of maximizing shareholder value took hold. Corporate executives who formerly lived merely like an especially comfortable middle class, and who gained prestige from sharing rents widely among corporate stakeholders, narrowed their focus to serving capital interests exclusively, and obtained compensation packages that tied their fates to that goal alone.

All of this might have made sense were it true that the only way to increase profits is to do things that add net value to the economy in which everyone else claims shares. But that’s the hard way to increase capital’s share of income, and thereby the income of top executives. It’s much easier for the top 1% to make money by creating and exploiting opportunities to gain at the expense of everyone else. Under the guise of ‘free’ markets, what was created was an alternative set of rules and practices rigged to serve capital owners and executives at the expense of ordinary workers, retirees, and young people. Let us count the ways.

  1. IP monopolies have been strengthened worldwide. So-called ‘free’ trade deals have replaced labor-protecting tariffs with steeply increased capital-protecting IP regulations. Copyright terms have been extended far beyond any credible incentive effects.

  2. Central banks across the OECD have practiced austerity, or failed to make unemployment reduction a priority, thereby gratuitously increasing unemployment to serve capital interests. Fiscal policy, too, has kept demand for labor weak, even while profits have soared. That r>g is due in part to g-depressing monetary and fiscal policies.

  3. Laws and regulations regarding credit and bankruptcy have been rewritten to favor creditors. In the U.S., bankruptcy no longer fully discharges personal debts for many people. Millions of college students in the U.S. labor under mountains of undischargeable student debt. Usurious payday and title loans reinforce the cycle of poverty for more millions. Many creditors’ business models are predatory, in which profits are generated by terms that trap people into spirals of debt, default, and accumulating fines and fees, and are deliberately designed to prevent people from paying off the loan, so they must pay interest and fees for a longer period. Regulators failed to reduce the principal owed on home loans after the financial crisis, gratuitously extending the length of the recession. In the EU, too, German-led monetary policy has strongly favored creditors over debtors, leading to recession and mass unemployment in the peripheral Eurozone countries.

  4. Antitrust enforcement has weakened, increasing the dominance of big firms that exploit their market power, fattening profits and executive compensation.

  5. Financial deregulation has driven capital away from growth-supporting investment, toward speculative trading that increases financial instability. It has also led to a diversion of talent and energy into negative value-added activities such as high-frequency trading, frontrunning, and LIBOR manipulation. The rise of banks ‘too big to fail’ has led to a culture of impunity and lawlessness in the financial industry. Notwithstanding massive fraud in the mortgage industry and serial criminality on the part of major banks such as J. P. Morgan, virtually no guilty bankers have been prosecuted for their roles in the financial crisis, and fines capture only a small fraction of profits from illegal dealings. All of this has increased inequality.

  6. On the labor side, in the U.S., basic employment laws are unenforced or carry penalties too low to deter, leading to massive wage and tip theft, forced work off the clock, and numerous other violations, especially at the low end of the wage scale. Employees are routinely misclassified as independent contractors, as a way to escape requirements to provide benefits, pay social insurance taxes, and fob business expenses onto workers. Young workers performing useful services for their employees are routinely misclassified as interns, so they don’t have to be paid at all.

  7. The rise of contingent and temporary labor and labor subcontracting has also enabled corporations to shed responsibilities for providing decent pay, benefits, and working conditions–a pure shift of income from labor to capital (or, for nonprofits such as universities, a pure shift of income from contingent workers such as adjunct faculty to the pockets of top-level administrators). Franchising performs similar functions, whereby the franchisor imposes costs and pricing structures on individual franchisees that all-but-guarantee that the latter cannot clear a profit without violating labor laws. Outsourcing abroad, including to enterprises that exploit forced and defrauded workers, magnifies these problems. These practices are due to a failure of employment law to close loopholes that empower firms to pretend that their employees are someone else’s responsibility.

  8. U.S. law has systematically failed to protect workers’ contractual pension rights. During stock booms, firms are permitted to skim supposedly excess profits in their pension funds for distribution to shareholders. In the inevitable bear market that follows, they dump now severely underfunded pension funds as hopelessly insolvent. Public pensions, too, have been underfunded or raided for decades.

  9. The shift from defined-benefit pensions to defined-contribution retirement plans has put the onus on naive investors to invest their savings. Yet financial advisors are free to peddle high-fee low-return investments to them, pretending to act in their interests, leading to returns on 401(k) plans for the ordinary investor that are well below r. While regulations have been proposed to end this practice in the U.S., its prevalence represents a pure shift of income and wealth from labor to capital, and from ordinary workers to high-paid financiers.

  10. In the U.S., labor laws protecting the right to organize have been violated with impunity at least since the 1980s. The decline of labor unions, in turn, has led to a decline in labor’s political influence for all policies affecting workers, whether they are unionized or not.

  11. In the U.S., the minimum wage has not kept up with inflation. Without the backstop of a minimum wage, much of the incidence of publicly provided benefits to low-wage workers, such as food stamps and the earned-income tax credit, accrues to major corporations, who don’t have to pay as high wages to induce the same labor supply.

From an ideological point of view, much of this can and has been peddled to the public as ‘free’ markets and ‘deregulation.’ The reality exposes the vacuity of these very ideas. In any advanced economy, the state must be involved in promulgating the constitutive rules of the economy. It can no more get out of the business of regulating the economy than the Commissioner of Baseball can get out of the business of promulgating the rules of Major League Baseball. The only real question is, in whose interests are the rules designed?

Ideology matters for politics. Once people have acquired income or wealth through the market, they feel strongly entitled to it. In the U.S. and increasingly in the rest of the OECD, the population at large, taken in by such representations, is reluctant to tax. Redistributing income and wealth by means of taxation, as Piketty proposes, becomes harder once people have it in their hands. We need to scrutinize the rules by which income and wealth get generated through the market, before it is taxed. They have been changing in a plutocratic direction for the past 45 years. The rule changes have not only increased r (at least for the top 1%), but also depressed g, by increasing monopoly power, shifting savings from real investment to speculation and scams, shifting top talent from production to value-extraction, and depressing aggregate demand.

Getting this story out is critical to changing politics. For plutocracy still must nod to what we might call ‘weak’ Rawlsianism: that inequality cannot be justified without showing that it delivers some benefits to the 99%. (It’s not for nothing that one of the leading arms of plutocracy is called the Club for Growth.) Exposing the ways the game is rigged, as Elizabeth Warren has been doing, should open more levers to change than focusing on taxes alone–levers that should also help limit the pace of increasing inequality by raising g.

{ 252 comments }

1

Rakesh Bhandari 12.07.15 at 5:50 pm

I think that this comments misses the force of Piketty’s inequality r>g. Even if intellectual property laws were weaker (and rents thereby smaller) and and labor laws and Keynesian policies stronger (and thereby g higher) and even if small savers had a bit higher rate of return, the Piketty inequality would still be in operation (and big savers would still have much relative returns), giving us the uncanny return of a rentier society.
In short, this comments seems to me to miss the force of the r>g inequality.

2

Dan Cole 12.07.15 at 6:12 pm

I don’t think it’s a case of either or. it’s true that Piketty does not pay sufficient attention to institutions and institutional changes that have increased returns to capital and reduced intergenerational economic mobility (up or down the ladder). His model stands on its own, but of course the returns to investment depend heavily on the kinds of institutions Prof. Anderson discusses.

I do think, however, she might add one to her list: the decline of the Rule Against Perpetuities (in the US) and the re-emergence of dynastic trusts, which lock in wealth across generations. This is still very much an institutional change in progress, state by state. Its probably will not begin to significantly affect social mobility statistics significantly for a generation.

3

Rakesh Bhandari 12.07.15 at 6:23 pm

I don’t agree that Piketty ignores institutional changes or changes in property laws, e.g. the end of human chattel, the operation of war commissions, the evolution of minimum wage laws, the weakening of collective bargaining, and the “stakeholder” nature of Rhenish capitalism. Piketty is aware of all this, but he still reasons on the basis of his r>g inequality that the return of a rentier society is most likely except through a global tax on wealth. The principal reason: even if through institutional reforms r is reduced a bit, g is likely to revert to historical averages even with the right Keynesian policies. That gives us a r>g inequality great enough to yield an actual rentier society (not the petit rentier one we now have).

4

Zamfir 12.07.15 at 6:34 pm

Excellent list, thank you.

5

An On 12.07.15 at 7:38 pm

I’m sympathetic to the direction of this argument, but does a list so focused on the US provide an adequate explanation of a global phenomenon?

6

chris arnade 12.07.15 at 8:08 pm

A wonderful list; which in many ways can be summarized (not to diminish the longer version) as various forms of, “The wealthy have been allowed to write rules that work best for them”

This has meant diminished rules, and the enforcement of rules, for Wall Street and large corporations, sold under the free market notion that individual liberty is collectively beneficial.

Yet at the same time increased and aggressive regulation has been applied to poorer folks (Broken Windows policing and the War on Drugs) under the theory that individual liberty can collectively be corrosive.

The latter hasn’t just been a moral outrage, it has also helped to devalue labor. It is harder to increase human capital when you are subjected to onerous rules and regulations.

7

T 12.07.15 at 8:32 pm

8

Robert 12.07.15 at 9:15 pm

Robert Reich’s new book, Saving Capitalism, is basically a longer statement of the thesis of this post, albeit not as a comment on Piketty. Krugman has a review in The New York Review of Books.

9

krippendorf 12.07.15 at 9:41 pm

see also work by Grusky and Weeden, who have been pushing the line that rents are ubiquitous throughout the labor market, and not just in the top 1%, for quite some time.

But, they are sociologists, so they are easily ignored by Stiglitz, Reich, and other economists who are trying to make rent-based arguments for rising income inequality.

10

Peter K. 12.07.15 at 9:53 pm

I am excited about this seminar and discussion.

What I believe Piketty has said is that *r* remains remarkably constant despite how one would think that the laws of supply and demand would mean that an oversupply of capital and slower growth would decrease *r* but it remains high because of a variety of factors.

Piketty discusses how Depression and Wars changed the dynamic so that inequality decreased in the post-war years. The list of policies here describe the many reasons why inequality has increased again and growth has slowed down.

As Jared Bernstein wrote “…since the late 1970s, we’ve been at full employment only 30 percent of the time (see the data note below for an explanation of how this is measured). For the three decades before that, the job market was at full employment 70 percent of the time.”

https://www.washingtonpost.com/posteverything/wp/2015/10/19/full-employment-a-bipartisan-goal-thats-missing-from-the-candidates-debates/

I believe macro policy (monetary, fiscal, trade) explains a large part of why growth has slowed and inequality increased since the 1970s.

Deregulatory “free market” policies pushed by rich financiers and conservatives have made the economy more volatile and prone to the boom-bust cycle. During the boom, most of the gains accrue to the top, and after the bust, macro policy has been insufficient to bring about a swift recovery, again exacerbating inequality.

11

Tabasco 12.07.15 at 10:38 pm

Central banks across the OECD have practiced austerity

For nearly a decade central banks across the OECD have kept their policy interest rates at or near zero. In a couple of countries they are less than zero.

Whatever this is, it is not austerity.

12

Rakesh Bhandari 12.07.15 at 10:57 pm

From the phone. Even if r on average falls from 6 to 4 pc. and growth goes from 1.5 to 2pc you will still get a rentier society. Remember also wealthy will have higher average return m, say in this ex 6 pc. Then you have wealth growing 3 x the rate of income. I earlier argued that we should distinguish P’s critique of rentier society from the Reich/Stiglitz critique of monopoly rents. P’s whole point is that rentier society arises out of even competitive capitalism

13

Rakesh Bhandari 12.07.15 at 11:01 pm

What piketty does not analyze is nature of ideological hegemony in a rentier society. I already pointed to Bukharin’s critique of rentier ideology in the economic theory of the leisure class. It makes for a fascinating comparison

14

Frank Wilhoit 12.07.15 at 11:35 pm

Chris Arnade @ 6:

A. “…the theory that individual liberty can collectively be corrosive….” is merely a cover for sadism.

B. “onerous rules and regulations” are only complained of by entities who do not wish to be held accountable.

15

bob mcmanus 12.07.15 at 11:44 pm

What piketty does not analyze is nature of ideological hegemony in a rentier society

As I read him, he doesn’t need to, because it’s irrelevant.

What Piketty’s numbers prove is that it wasn’t ideology or politics or unions or social movements and programs that gave us the Great Compression and decreased inequality but revolution, depression, and catastrophic war. Certainly history shows that every catastrophic war etc did not necessarily led to greater equality, but there is very little evidence for increases in equality without radical social disorder. Piketty explicitly says toward the end that moderate tax increases or redistributive social programs have had little effect, that the lower baseline after WWII was determining.

What is taken from the above can be up for discussion, perhaps the best can be done during peacetime is ameliorative efforts within a context of rising inequality, and ideology can help with those besides preparing for the inevitable collapse. But effective demand management will quickly fail for political reasons, see Kalecki.

If the above looks like Marxian praxis, it’s no coincidence. Piketty’s recommendation, taxing global wealth at confiscatory rates, should be understood as a practical recommendation. I think we all understand what it would take to tax away 40-50% of gross Saudi or Brunei or American wealth to distribute to Africa and South America, and Piketty surely was not unaware.

16

bob mcmanus 12.07.15 at 11:52 pm

I apologize, Piketty only asks, although I think he says “initially,” for a lower global wealth tax rate, perhaps a few percent to counter r>g. I will dig out my copy. I am not at all bothered if Piketty details no clear process to confiscating global wealth, he surely doesn’t have to, and it would get in the way of his message: that paths to equality are radically limited.

17

Peter T 12.07.15 at 11:54 pm

What struck me most about Piketty’s data was the near constancy of r over centuries. This suggests to me that it is not a matter of balancing supply and demand for capital, but a structural feature (perhaps the return necessary to sustain the hierarchy of production?). So r, in itself, has little to do with the distribution of return: it could be spread across a large middling to wealthy class, or concentrated in the 0.1%, as political factors dictated.

I will add that the underlying mental model in much of economics seems to be the gentleman’s estate. There is land, income for investment or consumption as prudence and virtue dictate, the lump sum in consols…with the ideal the improving landlord. That categories such as “capital” or “labour” do not stretch to countries does not seem to cross the imagination.

18

Mike Furlan 12.08.15 at 12:13 am

N. Taleb criticizes Piketty on mathematical grounds:

“What is worse, rejection of such theories also
ignored the size effect, by countering with data of a different
sample size, effectively making the dialogue on inequality
uninformational statistically.”
http://arxiv.org/pdf/1405.1791.pdf

I thought Piketty addressed this issue, for instance in looking at countries with relatively constant populations over time like France, and admitting that less could be learned from looking at a United States that grew from 3 to 300 million. But I’m sure that I’m missing many important aspects of this question.

19

Rakesh Bhandari 12.08.15 at 12:27 am

@18 If population is growing at a fast rate and therefore g as well, the fundamental inequality will be attenuated, and it will be difficult to see the long-term consequences of r>g in such a society. France gives us a better laboratory to see the likely effects of the fundamental inequality going forward than the US has hitherto provided.

20

Rakesh Bhandari 12.08.15 at 12:29 am

@15. I don’t think this is quite right. Piketty thinks a fundamental problem with rentier society is in fact ideological, viz. that it cannot be squared with the meritocratic values that provide normative support for competitive markets. This raises the question of what the elements of a rentier ideology are. The only one I know to have provided an answer is Bukharin.

21

Rakesh Bhandari 12.08.15 at 12:52 am

@17 raises difficult questions. Some economists have claimed that r>g is just what you would expect from dynamically efficient economy, but this needs to be spelled out. Piketty has a complex section which I have not yet fully understood on why r being positive, and greater than g, cannot be explained in terms of a psychological theory of time preference. Perhaps another way of thinking about this would be: what would happen if r were to fall below g? Would there be mechanisms to restore Piketty’s fundamental inequality? Piketty, I think, is saying “yes”. So I shall re-read that section to get a better understanding of his argument.

22

John Quiggin 12.08.15 at 1:01 am

“Central banks across the OECD have practiced austerity”

As Tabasco observes, this point is loosely phrased. Austerity is a fiscal policy, not a monetary policy. But central banks can enforce austerity by refusing to accommodate government budget deficits. The ECB has clearly done this (as it was set up to do). In other cases, governments and legislatures have imposed austerity, with the support of central banks.

The US Fed is one example where the central bank has been less supportive of austerity than the legislature.

23

ZM 12.08.15 at 1:25 am

A question for the more economically minded – Would one reason that r is reasonably stable over a considerable time frame be because it is determined not by the most wealthy holders of capital (who presumably could afford to take a lower rate of return on a long term basis), but by the less wealthy who depend on a higher rate of r as otherwise their smaller investments wouldn’t be financially rewarding?

For instance, in Australia workers have compulsory employer paid superannuation investments, and if the rate of return on these was lower I don’t know that the policy of moving people to self-funded retirement by superannuation as opposed to government pensions would be feasible?

24

Peter K. 12.08.15 at 1:37 am

“Central banks across the OECD have practiced austerity”

I think that is right. They’ve not supported quick recoveries and have been overly fearful of phantom inflation.

http://cepr.net/blogs/beat-the-press/paul-krugman-larry-summers-and-the-fed-s-unused-ammunition

25

Sebastian H 12.08.15 at 1:39 am

I’m very sympathetic to much of the list of bad policies. I have a question about this though: “On the ideological front, several theories served to rationalized policy shifts in favor of increasing capital shares and top labor incomes. The stagflation of the 1970s was successfully blamed on Keynesian economics, fiscal irresponsibility, a bloated welfare state, militant labor unions, state regulation of the economy, and supposedly incentive-destroying high marginal tax rates on capital incomes and the rich.”

What should it have been blamed on?

[I’m very open to the idea that the lessons of the 70s were overlearned–i.e. just because too much inflation is bad doesn’t mean we should worry about it when it is below 5%. But this comment suggests that something else entirely was going on]

26

Rakesh Bhandari 12.08.15 at 1:56 am

@25 in regards to the quote from Anderson who, along with many others, is missing in my opinion Piketty’s main thesis of how the normal operation of competitive or even social democratic capitalism or of course monopoly capitalism yields a rentier society in the absence of wealth and corporate taxation. Does not Piketty argue that the return to rentier society was underway before the Anglo-American neo-liberal turn (though it did obviously accelerate it) and has been happening even in societies not as neo-liberal as the Anglo-American ones? I read @5 as making this important point.

27

jake the antisoshul soshulist 12.08.15 at 1:56 am

I am not an economist, but the source of the blame was political. I have heard that the primary causes of “stagflation” were “printing money” to cover the debts from the Vietnam War, increased oil prices and supply issues driving up costs, and economic competition from Japan reducing demand. The economic elites saw an opportunity to
take advantage of the crisis and blame it on policies they did not like. And push to replace them with policies that were advantageous to those elites. Ronald Reagan was a very successful salesman for this.

28

Rakesh Bhandari 12.08.15 at 1:59 am

Maybe the American left is so focused on the critique of bad Republicans like Ronald Reagan and H.W. and W. Bush and so excited about Warren and Sanders–these political choices setting the limits of theoretical analysis– that it cannot countenance Piketty’s deeper structural critique?

29

Peter T 12.08.15 at 2:29 am

Piketty uses “capital” or “wealth” to refer to any asset which provides a stream of income. This is, in my view, correct. This is quite distinct from “capital” in the ordinary economic sense. Much – in fact most – economic capital does not yield income (roads, schools, food crops…) and so does not count. There is no reason to believe that wealth in the first sense does or should correspond to capital in the second sense. Our collective capital dwarfs wealth, while much wealth is simply extractive. To count a claim on tax revenues (a government bond) in the same class as a terraced field is a major mistake.

From the late C19 on much private “capital” was withdrawn into the public spheres (eg private tolls or offices), a move that accelerated in the wars. Since 1980 this trend has reversed. Discussion of the amplitude or scarcity of capital should note that it is a legal and political category subject to large arbitrary changes.

30

Bruce Wilder 12.08.15 at 2:56 am

Piketty’s deeper structural critique

Does Piketty have one? ‘Cause then I missed it. It seems to me that Piketty is presenting the challenge of facts. He takes care to outline how the facts he documents are logically related, as in the analysis of how changes in the share distribution of income (between labor and capital) relates to economic growth and to the value of accumulated wealth as a stock. That’s not “deep” or structural, though it is certainly necessary if we are to understand the facts as facts.

The first striking thing to me in Piketty’s work is what Peter T discusses above at 2:29 am (@ 29): the distinction between wealth and capital, confusion about which powers the ideologies of more than a few economists and others. Just maintaining that distinction, while discussing the wild swings in the share of income going to capital over long periods of time forces attention to the politics. Is that “structure”?

31

notsneaky 12.08.15 at 4:06 am

@Rakesh – but the whole r vs. g thing in Piketty, while central to his book, is also the part that makes the least sense. It’s made up, theoretically unsound, and with no evidence to back it up. It’s junk. An accounting relationship is not a “law”.

And it’s really Piketty’s single minded focus, based presumably on his desire to provide a grand “one size fits all” explanation for the phenomenon he’s discussing, which leads him to sideline all the possible institutional explanations, such as the ones enumerated above (not that I agree with all or even most of them)

32

notsneaky 12.08.15 at 4:18 am

In terms of r vs. g

In linear production model – r > or or or g always but capital’s share in income is constant. Taxing capital doesn’t matter for distribution.

In Ramsey model with endogenous saving and non-Cobb Douglas production – r > g but same criticisms as above imply.

The only one story about r vs. g out of the whole book which sort of makes sense is that if r > g then capital income can become more unevenly distributed (even as capital total’s share stays constant). But even that is based on some sketchy assumptions and relaxing these even slightly can completely flip the result.

The Journal of Economic Perspectives V 29/ N 1, 2015 has a symposium on the topic and it pretty much consists of various polite ways of saying “good data, but the r vs. g thing is nonsense”

33

Omega Centauri 12.08.15 at 4:39 am

I’ve argued before that the conclusion that the simple inequality r>g leads to unlimited inequality is wrong -or at best incomplete. There are multiple ways that concentrations of capital can be, and are dissipated (i.e. broken up into smaller bits owned by more people). Taxes, and “death-taxes” is only one mechanism to accomplish this. Having on average more than one inheritor is another. Think for example of the Saudi Royal family, which controls great deal of the wealth of the Kingdom. It isn’t all concentrated in one nuclear family, there are now thousands of princes, after not too many more generations a plurality of the country will be able to claim royal inheritance. Also there are other mechanisms, that can dissipate wealth concentrations, including luxury goods: Maserati is distributing some wealth from the super rich, to its shareholders and employees… Major donations to charity is any other. Still another comes from the application of the saying “a fool and his money are soon parted”: some of the progeny, will be separated from their inheritances. One can’t just use a simple theory of the evolution of the distribution of wealth and income, if you ignore wealth dissipation effects you will get a wildly wrong result.

So in order to control or reverse the tendency towards ever increasing inequality, there could be deployed multiple strategies, all of which are aimed at increasing the dissipation of concentrations of wealth.

34

Omega Centauri 12.08.15 at 4:46 am

I also think there exist mechanisms in the economy which tend to stabilize R. The most obvious is that there are only so many profitable investments available at any given time, and supply/demand effects should lead to lower R if the amount of available capital gets too high or to increase it if there is less capital than investment opportunity. Also the tendency to
spend wealth on immediate consumption versus investment changes as the expected return on investment changes. These effects should usually lead towards returning R towards some long term sustainable value.

35

Peter T 12.08.15 at 4:49 am

notsneaky

My reading of this is:

consider r as the average rate of return on the things politics allows one to generate income from owing (land, slaves, government offices, army commissions, tax revenues, corporate stocks…). Piketty finds this to be a near constant 4% or so.

Piketty observes that long-run growth is generally less than this, that some part of r is plowed back into further wealth acquisition and that larger wealth-holdings earn higher returns and are, on the whole, more immune to negative shocks. So wealth concentrates, and takes a larger proportion of overall income until there is a re-set. This can be revolution, war, aggressive state redistribution or a redefinition of what is allowed to count as wealth.

It’s neither a law nor an accounting identity. It’s a historical trend.

36

notsneaky 12.08.15 at 5:34 am

The fallacy occurs in the phrase “So wealth concentrates…”. This doesn’t follow. You can have the historical trend that r is greater than g and have anything you like happen to capital’s share. You can even have r greater than g and have anything you like happen to concentration of capital ownership.

And Piketty does repeatedly refer to this as a “law”. Which is really annoying. And obnoxious and presumptuous.

37

Robert 12.08.15 at 5:49 am

My name links to my demonstration that Piketty is mistaken about his theory about the consequences of r exceeding g.

38

Peter T 12.08.15 at 6:10 am

“You can have the historical trend that r is greater than g and have anything you like happen to capital’s share”. Well. you can. But you don’t (without politics).

39

notsneaky 12.08.15 at 6:20 am

But that’s the point though. It’s not about r vs g. It’s about politics. r vs g could be doing whatever it is r vs g wants to do and capital’s share could be doing its own thing. So if we see capital’s share going up we shouldn’t say “it’s r vs g”. We should say “it’s politics”

40

Daniel Frick 12.08.15 at 6:31 am

Note the names of the Fortune 500. Now cross-check vs the names of the robber barons in turn of the 20th century in any high school history book. If Pinkety was right, most of the names should be the same. In theory, theoretical societies work the same as real societies.
In theory.

41

Peter T 12.08.15 at 6:44 am

No. Because the trend is there, inherent in the system. The politics is about limiting/reversing it. Unless, by “politics” you mean the structure of society.

42

notsneaky 12.08.15 at 6:48 am

What trend and how is it “inherent”?

43

notsneaky 12.08.15 at 6:53 am

If by “trend” you mean “r is greater than g” then that is neither a well established trend nor is it in any way (“inherently”) related to the degree of inequality. Not empirically. And theoretically, Piketty does not actually explain how it would be.

44

Bruce Wilder 12.08.15 at 7:25 am

Omega Centauri: . . . there are only so many profitable investments available at any given time, and supply/demand effects should lead to lower R if the amount of available capital gets too high or to increase it if there is less capital than investment opportunity.

You think there’s a fixed set of “investment opportunities”? And, investing in those opportunities is subject to some kind of satiation limit?

What if the “investment opportunity” is opening up publicly funded charter schools and taking a cut of public spending on education? What if the “investment opportunity” is closing community banks and opening up payday lenders in some neighborhood? What if the “investment opportunity” is to close a union bakery, stiff the workers on pensions, take the brand-name and open up a non-union shop somewhere else that pays the bakers left and the executives more? Are those opportunities diminished by the wealth invested in them?

45

Dipper 12.08.15 at 7:49 am

point 5 looks a bit weak.

– point 5 starts off with speculation and ends with illegality. They are separate. What exactly is your point here?
fines capture only a small fraction of profits from illegal dealings where is your evidence for that?
– Where is your evidence that the amounts of money diverted in the scandals you mention are big enough to have resulted in measurable increases in global inequality?
virtually no guilty bankers have been prosecuted for their roles in the financial crisis Tom Hayes has just been sentenced to 14 years, although he has an appeal. Your assumption that you know who is guilty without the need to go through a trial sounds like a witch hunt.

Why no mention of banks roles in Mergers and Acquisitions? These have reshaped ownership and control of many industries without any apparent benefit.

Why no analysis of regulation post-crash? Significant changes have taken place in Financial Markets, many of which left-leaning people should support. Are they working?

46

Peter T 12.08.15 at 7:50 am

notsneaky

I fail to see how this is hard.

If some wealth-generating resource (say, land) is initially evenly distributed and

1. claims on the returns can be traded then
2. ownership will gradually concentrate as random shocks eliminate smaller shares (the same process that has eliminated pretty much all but 100 Chinese surnames).
3. This trend will be emphasised if the owners of larger shares re-invest more and/or if they enjoy higher returns.

So you end where 7,000 closely-related families owned 80 per cent of the land in Britain.

All these things happen to be true. It’s not a specifically capitalist phenomenon; it happens wherever the claim to income is not directly connected to the generation of it (that is, wherever there is a socio-economic hierarchy).

One function of politics is to prevent this tendency running to extremes, by confiscation, redistribution or redefining what can be claimed against. Where politics fails at this, the productive machine breaks down and the hierarchy scales back.

47

Brett 12.08.15 at 8:11 am

Omega Centauri brought up a point with the Saudi royal family. Wealthy families tend to burn out most of their money in a few generations at most – very, very few of them remain wealthy for longer than that. How does that work in a system supposedly dominated (or becoming dominated) by patrimonial capitalism? Just intra-elite competition in zero sum terms?

48

Brett 12.08.15 at 8:17 am

Sorry, thought about this a bit more-

Financial deregulation has driven capital away from growth-supporting investment, toward speculative trading that increases financial instability.

It’s more that they’re moving money out of companies with low returns on investment to higher investments, which usually translates to investments in risky assets or equities (hence the boom in share prices). If you want an idea of what the view-point of a supporter for this is like, here’s a post that claims that the auto companies in the US (which did significant share buybacks) had essentially negative returns on investment for a while. If you’re an investor, you want to pull your money out of such a company ASAP and move it somewhere else – which the owners and managers of auto companies obviously don’t want.

Of course, that’s from the perspective of investors. If you’re a worker in such a firm with no ownership stake in the companies, just work for wages, then who cares what happens to the share value of the company? What matters is whether the company at least breaks even, and better yet prospers and distributes more of that revenue as wages.

49

reason 12.08.15 at 8:47 am

Brett @47
” Wealthy families tend to burn out most of their money in a few generations at most “
That was true several decades ago, but is it still true?

50

reason 12.08.15 at 9:01 am

It seems to me that the argument between Peter T and notsneaky is misdirection from both sides, because the implicit definition of ‘r’ that they are using is not the same.

51

TM 12.08.15 at 9:50 am

17: “What struck me most about Piketty’s data was the near constancy of r over centuries.”

That struck me, too. In a dynamic, ever changing universe, Piketty’s claim to have found something akin to a constant of capitalism – without naming any mechanism that would explain this one aspect of capitalist economy to have stayed near constant for centuries, while everything else has constantly changed – should induce a healthy dose of skepticism.

I’m rereading an earlier thread on r>g (http://crookedtimber.org/2014/10/13/r-g/) and my impression (you can read my detailed arguments there) remains that hardly anybody is actually taking Piketty’s r>g theory seriously, in the sense of a scientific model supposed to describe reality and capable of being tested empirically. As I have shown, as a model (rather than a slogan) r>g doesn’t pass the reality test. It’s not consistent with the data Piketty himself presents. For example, Piketty reports r minus g of about 3.5% during most of the 19th century, which would imply a roughly 32-fold increase in the capital-income ratio over a century, yet he reports (http://piketty.pse.ens.fr/files/capitalisback/F3) that capital income ratio in the UK was *constant* at about 700% throughout the 18th and 19th centuries.

52

Ebenezer Scrooge 12.08.15 at 10:50 am

@44
Tom Hayes’ conviction is irrelevant to Elizabeth Anderson’s point. I quote her: “virtually no guilty bankers have been prosecuted for their roles in the financial crisis”. Lots of bankers have been prosecuted, mostly for insider trading. None of them, except for one midsized fish in Florida whose name escapes me, were prosecuted for their role in the crash of ’08. That includes Hayes, whose acts may have been morally heinous, but had almost no effect on financial stability.
Then again, I’m not sure I agree with Anderson’s tacit assumption that banker criminality caused the crash of ’08, at least if one defines criminality in terms of the criminal laws that actually exist. Many bad acts are not crimes, and many of those that are crimes are not prosecutable in a legal system that formally favors defendants, and practically favors well-heeled defendants.

53

Dipper 12.08.15 at 11:04 am

@52

Well yes, agreed, but Tom Hayes was convicted for LIBOR rigging. If that is not relevant to Elizabeth Anderson’s point why is LIBOR quoted in point 5?

Banking provides a vast and rich banquet of scandal, wrongdoing, and inappropriate behaviours. But if you want to make serious points you can’t just fill your plate with whatever is in front of you. You have to pick carefully, and join the dots with suitable evidence. I don’t see much of that in point 5.

54

Peter T 12.08.15 at 11:11 am

reason @ 50

Maybe. I gave a definition of r at 35. My copy of Piketty is on loan, but I think it’s consonant with his definition – the return on wealth.

re TM @ 51 – a good deal of wealth was legislated away in the C19 (mostly private rights). A further large lump vanished when agricultural land values crashed with cheap bulk shipping arrived, and Irish land values were earlier affected by the legislated end of rack renting and, of course, the Famine. So we have the classic trio of politics, Malthusian checks and technology. I don’t know if these together explain why the wealth/labour proportion did not rise as much as might be expected, and IIRC Piketty does not go into that level of detail.

55

TM 12.08.15 at 11:36 am

Peter T: these factors are in direct contradiction to Piketty’s claim that the capital-income ratio remained constant from at least 1700 to 1900.

“I don’t know if these together explain why the wealth/labour proportion did not rise as much as might be expected” – if you are referring to the wealth-income ratio, it didn’t rise *at all* according to Piketty.

56

MisterMr 12.08.15 at 12:25 pm

[Previously commented as Random Lurker]
Disclaimer: I’m not an economist, just a guy who reads way too much Marx on internet.

I’ve been thinking about Piketty for a while, and I reralized that Piketty’s story sounds very similar to the theory of the 19th century socialist economist Rodberthus[1]. Rodberthus was a Labor Theory of Value (LTV) guy and a competitor of Marx on the market for socialist economists, so that there is a large section in Marx’s Theories of Surplus Value debating R.’s theories[2]. I also read Sraffa and I came to the conclusion that Sraffa’s model is equivalent to Marx’s LTV[3]. So here I go with my Marx-Rodberthus explanation of Piketty:

Piketty shows that there has been a secular decrease of the wage share in developed economies, and a more than proportional increase in wealth inequality. He explains this with his “r>g”, and with the hypothesis that there is a secular growth in the productivity of capital relative to the productivity of labor[4]. P.’s explanation has various problems:

1) The secular fall in the wage share just doesn’t fit well with the neoclassical theory that P. uses to explain it, expecially since it is quite evident that it depends on political choices; I’ll try to put it in a LTV context, that fits better the idea of a non technologically determined wage share.

2) P. treats all increases in wealth as increases of capital, however it is evident from his discussion of pre WW1 economies that most of this wealth is just the high value of land, that is a problem of asset appreciation not of capital accumulation.

3) P. has a more or less constant return on capital of 4%/year, that is just weird.

4) r>g just doesn’t make much sense.

Let’s start with the first point, with a LTV/Sraffian theory of distribuition. In this model, we have an aggregate capital K, and an output O, that have a fixed technological relationships, for example a K of 200 gives an O of 100 every year. This output of 100 is then distributed between workers and owners, so that if we have a wage share of 70%, 70 go to the workers, 30 to the owners as “surplus”, and the average rate of profit is 30/(200+70)=11%, while the return on capital is 30/200= 15%. Market prices then adjust so that every owner’s rate of profit is more or less 11%[5].
If we see this from an LTV point of view, supposing that a k of 200 needs 10h of labor to work, we have that 1h of labor produces an output of 10 (so the “real” value of 1h of labor is 10), but the wage of 1h of labor is only 7, leaving 3 as surplus value. Each 1h of labor mobilizes a K of 20, that is 2h of “value”[6].
However, owners will have an expected rate of return on capital (or they will “discount the future” at a certain level), and they will reinvest their profits as long as the actual returns are higer than their warranted rate of return. But K/O is a technologically determined constant, and the wage share is determined by policy, so where will this money go?
This is the “Rodberthus” part of the argument: in every economy there will be some “bottlenecks”, in the 19th century it was land, today it’s most likely distribuition chains, so the excess surplus value will be used to purchase those bottlenecks and therefore to raise the value of such assets, that become the capitalized value of the cashflows they produce at the “warranted” profit rate. Thus those rentier assets become a sort of pseudo-capital[7][8]. Hence the total wealth C will be higer than K, as it will be equal to K + pseudo capital[9].
Note that this explanation works both for the constant returns on wealth (C) of 4%, that is just the warranted rate of return and thus has nothing to do with technology, and for the appreciation of land and other rentier assets, as when the wage share falls, the difference between the rate of surplus and the warranted rate of return increases, so that the value of pseudo capital shoots up.
Finally, regarding r>g, this is just natural in a LTV model, because when productivity increases (g) the “real” value of 1h of labor increases, thus pushing up both profits and the “real” value of capital goods.

____NOTES____
[1] R.’s theory goes like this: the value of stuff depends on the quantity of labor needed to produce it, however workers get only a share of the total value that they produce, the remaining share (the surplus) is divided between profits from industrial capital and land rents (that for the classicals were two differnt things). The worker’s share depends on customs, and so does the surplus, however as the economy grows there are more and more industries, while the land is always the same, so that the share of surplus that goes to land rents increases at the expense of the share of surplus that goes into industrial profits. Also, the value of industrial capital depends for the LTV on the amount of labor needed to produce said industrial capital, but land has not a labor value, its value is just the capitalization of the cashflow of rents at the prevalent rate of profit, hence as the share of rents increases the value of land shoots up, creating a wealth effect (an accumulation of wealth).

[2] R. actually believed that M. plagiarized him, however after reading both I don’t think it’s true, but I don’t know what R. exactly believed to be plagiarized.

[3] This is disputed but Sraffa actually agreed with me. The point is that in M. theory, there is an amount of capital that is residual to total income, so that there is an “aggregate constant capital” and an economywide “organic composition of capital”. M. however explained this very poorly so that many readers (including outstanding economists) take the organic composition of capital to be just something relative to the single business, so that the whole economy is just constituted by “live labor” as it was in Ricardo. If we take in account “aggregate constant capital” M.’s model becomes the same as Sraffa’s model, only with sketchier math.

[4] This hypothesis is necessarious to square the fall in the wage share with the neoclassical theory of distribuition.

[5] This is the point of Sraffa’s “Production of commodities through commodities”, and is actually quite complex. My K/O constant is Sraffa’s R.

[6] In other words the K/O constant is Marx’s “organic composition of capital” where K is capital in Marx’s sense, that is produced means of production, that embody a certain quantity of “dead labor”. (Strictly speaking M. organic composition is K/wages, not K/O, but I think the point is clearer if we think in terms of K/O).

[7] In Marx’s sense that they don’t really embody “dead labor”, so they are just treated as capital, but they aren’t really.

[8] Also I think that overcapacity might be a form of pseudo-capital in a low wage share situation.

[9] As the wage share falls, the amount of wealth relative to income increases, as total wealth in the end is just the capitalized value of surplus at the warranted rate of profits, and as people can bet on the appreciation of capital assets, the economy also becomes more financialized and speculative.

57

Mike 12.08.15 at 1:16 pm

I have nothing to say about r>g. But something to say about “weak rawlsianism,” which is an interesting concept. What does the 99% get from the plutocracy? You can look at the Presidential primaries to see that: a promise of “safety” (and fear-mongering to make the general public feel that they “need” safety); a promise of “jobs” (and fear-mongering about minorities and immigrants who might “take” job). In neither case is there any serious thought about what “safety” might mean, or about whether “jobs” are meaningless if the pay is below poverty level. There’s no thought about whether proposed policies about either security or jobs might actually make the situation worse. If the situation becomes worse, that’s even better for the plutocrats, as long as the problem can be blamed on a minority.

“You hate what we hate; isn’t that enough?” It remains to be seen how well that will play out over the long term. Probably not well.

58

Omega Centauri 12.08.15 at 2:29 pm

Bruce Wilder: “You think there’s a fixed set of “investment opportunities”?”
Not a hard limit, but definitely there will always be a spread of investment opportunities with different expected returns. If investors are rational and well informed, they will invest in the most remunerative projects and leave the others starved for funds. If there is more investment money available, it has to choose from the remaining lower return projects, hence R tends to decrease as available investment capital increases. This isn’t to say that public policy (politics) couldn’t be used to change the distribution of investment returns, but for a given political/legal environment there are only so many ways to make money.

Sure the pool of available “projects” can be increased by applied imagination, and disregard of social ethics etc. But the general principal still applies. There are only so many street corners to set up three card Monte tables on, the best sites are taken early on.

59

notsneaky 12.08.15 at 2:53 pm

Peter, first we need to clarify what precisely we’re talking about (one of the annoying things about Capital20 is that Piketty never does this and then switches back and forth between different definitions of “inequality” committing various false equivocations and other fallacies in the process).

If by “inequality” we mean capital (or land’s) share in total income then in your example there’s nothing which actually affects what happens to it. That share could go up, could go down or it could stay constant. r vs g doesn’t explain anything about it.

If by “inequality” we mean concentration of ownership of capital assets then it’s possible that the process as you describe leads to higher concentration of wealth via “shocks which eliminate smaller” claims. BUT. The last part is not necessarily correct. It’s the fact that there’s random shocks which leads to capital ownership concentration, not anything to do with r vs g. Higher returns (higher r) may or may not affect how this process unfolds. Under some very implausible assumptions (both r and g are divinely determined) it could wind up like that. But as soon as you realize that r and g are actually related to each other, the logic breaks down.

There’s absolutely no explanation in Piketty’s book about how r greater than g should actually, logically, lead to higher inequality. There’s an assertion and some sketchy accounting. There’s a lot of claims about “inescapable logic of r vs. g” but no explanation of that inescapable logic. There’s a lot of discussion as to whether r is actually greater than g. But no explicit connection.

Empirically, we know that inequality – both in terms of capital’s share and in terms of capital ownership concentration – has increased. For some portion of the period, maybe r was greater than g (very debatable actually). So at best you have an empirical (very weak) correlation between a time of increasing inequality and r greater than g. But without an explanation of a casual link, this explanation fails. There’s a million other explanations/correlations that could be posited.

The implication is that there’s nothing “inherent” or “natural” about the increase in inequality that we’ve been observing. It was simply a consequence of policy choices which more or less redistributed income from labor to capital, and from small scale capital owners to large scale capital owners. We could very well have had r greater than g, a different set of policies and a different level of inequality.

So in general, the OP is correct here. Piketty is focused and determined to establish this r vs g thing, which isn’t even logically sound, that he misses the obvious roles of policy.

Here is Chuck Jones working through the logic of r vs g

http://pubs.aeaweb.org/doi/pdfplus/10.1257/jep.29.1.29

Here is Acemoglu and Robinson addressing both the theoretical and empirical weaknesses of the r vs g story

http://pubs.aeaweb.org/doi/pdfplus/10.1257/jep.29.1.3

60

notsneaky 12.08.15 at 3:00 pm

@17 and @51

The general, rough, constancy of r isn’t Piketty’s discovery. It’s one of the so-called “Kaldor stylized facts” that goes back to at least the 1930’s and even then Kaldor didn’t discover it but was just stating what was already known. Basically any economy (whether capitalist or not) which has a “steady state” will wind up with a constant r. (Malthusian economies might be exceptions)

61

jake the antisoshul soshulist 12.08.15 at 3:01 pm

I see an interesting discussion here about politics. It seems to me that neo-classical economics tried to Korzybskize political economy. At least in the sense proposed in The Manhood of Humanity. Which was to empiricize social science (which Korzybski referred as the pseudo-sciences. Ironic since Korzybski is pretty much now considered pseudo-science).
I do think it tells us something that they dropped the political part of the equation.
I suspose there was some intellectual honesty involved in that they were looking for
laws of economics outside of politics. However, I am of the opinion that everything involving human behavior is “political”. There is no “inverse square law” or “second law of thermodynamics” in human behavior. So I think making an empirical science of
political economy (or any other social science) is doomed to failure.

62

SamChevre 12.08.15 at 3:18 pm

Peter T @ 45

So you end where 7,000 closely-related families owned 80 per cent of the land in Britain.

That’s very very heavily dependent on primogeniture, which keeps the estates together.

Otherwise, landholdings would split into smaller and smaller estates as they were inherited by all descendants. This offsets the concentration effects at least somewhat.

There’s a reason that primogeniture and perpetuities were forbidden by American law a century ago. (The heavily-endowed non-profits are classic perpetuities.)

63

Lee A. Arnold 12.08.15 at 3:52 pm

As I understand it, r>g is an empirical law, not a theoretical law, yet several commenters here seem to be making this mistake.

As to what causes it? I would guess that some of it is unintentional, but most of it is intentional, one way or another, and thus includes politics. Not merely includes politics, but animates much of politics, really.

Elizabeth Anderson’s list of political rent-seeking is a good one, and a keeper. But for me, her “money quotes” come at the end, after an implicit question: What to do about it?

Anderson observes, “…much of this has been peddled to the public…they feel strongly entitled to [their acquisitions]…reluctant to tax…need to scrutinize…getting this story out is critical…exposing the ways…”

Yet when we attempt to push further down this avenue, we crash into another empirical law:

Attempts to “get the story out”, to “expose the ways”, crash into a very potent combination of the following factors: mental inertia and cognitive limitations; personal time constraints on learning new information; the extreme informational complexity of the proper description; —

— and now, a rather new development: the explosive proliferation of new info-media, which has led to the complete disappearance of a common reference narrative to which people know that they can refer, for information that is reasonably true. Who is going where, to find out what? Our whole system is running on in-group narratives.

In short, we are already given a democracy, more or less, and thus we could fix things, if voters knew the facts as opposed to being driven by their emotions. Our real problem is an enormous RHETORICAL problem, with various facets to it. (I think there is a way to go here, and you can click on my name for a possible answer.)

However, I want to make a very different point, about where to go:

At the same time as we have exacerbated the era of rent-seeking, we appear to be moving into an era of post-scarcity and technological unemployment. I don’t think we should see it as a bad thing, although people often appear to be confused by two faulty objections: 1. that focus on technological unemployment puts aside the rent-seeking explanation for inequality (no, it doesn’t); and 2. that technological growth must cause environmental damage (no, it needn’t).

We should instead incorporate it into the way forward.

Therefore, I don’t think it will help anybody to fight the increasing capital share — to “raise g”, in other words — only to maintain the 30-40 hour work-week for everybody, and to continue with this ridiculous system of the phony scarcity of money administered by interest rates.

Instead we should head towards solutions that comport with the real abundance: 10 hour work weeks for everybody (then, they might have the time to learn what is happening to them!) and print the money for universal goods and services which have supply-and-demand characteristics that do not automatically cause inflation (e.g., education, healthcare, retirement security). There is no sense in coursing everything through the private financial system, just so they can take their share for no value-added. We will end up right back here, where we are.

64

Lee A. Arnold 12.08.15 at 4:08 pm

The way to fight an empirical law, in other words, may be to change the empirics.

65

jake the antisoshul soshulist 12.08.15 at 4:22 pm

I don’t really understand the criticism of r>g. If I understand correctly some are saying that P. is claiming that r>g when g is sometimes apples, and sometime oranges.
Some seem to be saying that any relationship between r & g is nonsense.
And some seem to be saying that P. is misrepresenting or misreading historical trends.

Are any or all of these criticisms legitimate.

66

que_es 12.08.15 at 4:49 pm

Dan Cole at 2.

The repeal of RAP is only a small part of how laws are evolving to entrench capital in the trust law area. Spendthrift trust laws ensure that great wealth is exempt from creditors’ claims. Thus gifted wealth is “special” wealth. Not satisfied with exemption from ordinary contract creditors’ claims, the trend is not to allow even involuntary intentional tort victims access to a tortfeasor beneficiary’s spendthrift trust assets. In the meantime, new trust decanting rules direct courts to engage in estate planning for long-dead trust settlors, “renewing” their trusts by taking into account new circumstances and more favorable laws for the protection of their former assets, now benefiting their offspring. Changes to the rules involving equitable deviation of trusts also directs courts to permit cutting-edge estate planning by proxy for dead settlors, so long as it is of benefit to the donee/beneficiaries. I could go on . . .

67

TM 12.08.15 at 4:53 pm

notsneaky 61:

“Basically any economy (whether capitalist or not) which has a “steady state” will wind up with a constant r. (Malthusian economies might be exceptions)”

Now you are astonishing me. There are no “natural constants” in the economy and anybody who claims having found one is almost certainly a crank. And also, logically you cannot have a ‘steady state’ while capital accumulate faster than economic growth. Capital depreciation will end up swallowing the whole of GDP.

LAA 64: Piketty claims r>g as an empirical fact (or at least tendency) AND makes claims about the implications of r>g for inequality. Even though the latter aren’t really fleshed out as an explicit theory (as notsneaky says at 60), Piketty clearly claims (without demonstrating) some generizable correlation between inequality and the difference r minus g.

68

Lee A. Arnold 12.08.15 at 5:17 pm

Ze K #67: “…what’s the meaning of his suggestion to introduce an entirely unrealistic, politically, world-wide wealth tax?”

I can imagine that if one country were to “reverse the policies listed in the post”, then financial capital would flee across its borders, causing domestic economic unrest due to lack of real investment+ onerous deleveraging, until the population cries “Uncle” and the rent-seeking policies are re-enacted.

It could also be that the long duration of the empirical law + the dour human psychology portrayed in his examples from novels, have led Piketty to suspect that only a global institution will change individual human preferences in a lasting manner.

I tend to think that both of these are correct.

69

Lee A. Arnold 12.08.15 at 5:35 pm

TM #69: “Piketty clearly claims (without demonstrating) some generalizable correlation between inequality and the difference r minus g.”

I don’t understand where you are going. Is your argument that Piketty thinks that the correlation between inequality and r>g is theoretical?

70

Rakesh Bhandari 12.08.15 at 5:56 pm

Acemoglu and Robinson have argued that correlation between fundamental inequality (r>g) and inequality is weak on the basis of a statistical analysis. Piketty has responded that the effect he is focused on (growing share of capital income) is obscured when most of the inequality is among labor incomes; that the data we have on wealth quality is quite limited; but that what data he had points to a strong correlation as long as we take account of time lags and observe the process over 30 to 50 years.

71

notsneaky 12.08.15 at 6:23 pm

@65 “The way to fight an empirical law, in other words, may be to change the empirics.”

In that case it wouldn’t really be a law, would it?

(and anyway, it’s not even an empirical regularity. Sometimes r greater than g, sometimes g greater than r. Sometimes inequality goes up when r greater than g, sometimes it goes down. The correlation is weak)

72

notsneaky 12.08.15 at 6:25 pm

@66 “Some seem to be saying that any relationship between r & g is nonsense.”

Not quite. The relationship between the difference r minus g AND inequality is (more or less) nonsense.

In fact, it’s exactly because r and g can be related to each other that the difference between these has no implications for inequality.

73

notsneaky 12.08.15 at 6:30 pm

“Now you are astonishing me. There are no “natural constants” in the economy and anybody who claims having found one is almost certainly a crank. And also, logically you cannot have a ‘steady state’ while capital accumulate faster than economic growth. Capital depreciation will end up swallowing the whole of GDP.”

Ehh. I don’t know if there are “natural constants”. All I said is that if an economy has a steady state (where output, capital, consumption, etc. all grow at constant rates) then it will have a constant r. That’s pretty much a definition of a steady state. Whether that constant is determined by time preference, tax policy, institutions, politics, technology or God is up for debate, but it’d still be constant.

By “accumulates faster than economic growth” I assume you mean that capital grows faster than GDP. That would indeed not be a steady state. That’s why I said *IF* an economy has a steady state.

I don’t know what “Capital depreciation will end up swallowing the whole of GDP” means.

74

notsneaky 12.08.15 at 6:32 pm

@72 yeah I don’t buy it. Btw, do you have a link to Piketty’s reply? I had it at one point but it got lost. The part I’m confused about is how, if he’s focused on capital’s share and that’s his measure of inequality, will changes in labor income inequality affect overall capital share.

The point remains that any statistical relationship between r-g and inequality (of either kind) is not really there unless you really massage the data.

75

dsquared 12.08.15 at 7:08 pm

Basically any economy (whether capitalist or not) which has a “steady state”

Like France, say?

The word “economy” in this sentence is really dangerously ambiguous in whether it refers to an actual economy (in which case it’s talking hypothetically about something which has never been observed and probably could never be), or to a mathematical entity (in which case it’s necessarily true). This isn’t just a problem of whether the model is accurate enough to be useful – the idea that this equilibrium property is relevant to Piketty’s book is an empirical claim (IMO an extremely strong and definitely incorrect one).

76

Lee A. Arnold 12.08.15 at 7:12 pm

Notsneaky #73: “In that case it wouldn’t really be a law, would it?”

Not in the physical sciences. But what about in the social sciences?

77

notsneaky 12.08.15 at 7:18 pm

” the idea that this equilibrium property is relevant to Piketty’s book is an empirical claim”

I guess it’s empirical in the sense that we could open up Piketty’s book and check if he talks about economies and their steady states in there.

(or in other words, if you don’t like this approach, talk to Piketty)

78

notsneaky 12.08.15 at 7:20 pm

“Not in the physical sciences. But what about in the social sciences?”

I actually tend to agree with the critics of economics in that we (social scientists) really shouldn’t go around making claims about “laws” in general. But even by the standards of social science “laws” (say Gresham/Copernicus’ Law or the Iron Law of Oligarchy) it doesn’t do to well.

79

TM 12.08.15 at 7:46 pm

75: “All I said is that if an economy has a steady state (where output, capital, consumption, etc. all grow at constant rates) then it will have a constant r.”

Sure, if your definition of steady state includes r constant, then any economy at steady state will have r constant. That’s a tautology.

80

R Richard Schweitzer 12.08.15 at 10:05 pm

Perhaps we can hope for something more than a rehash of the same old, same old.
Perhaps some work on the classifications of “Capital;” the productive, the intermediary, the depleting, depreciating; the obsolescence bound, etc.
Perhaps some work on the differences in the *functional* returns (if any) to each type of classification. Functional would include rates of increases in costs of replacement, changes in costs of incurred obsolescence.

Perhaps some work on non-income forms of distribution (such as political transfers of costs) and the modes (other than measured incomes) of sharing participations in each “economy.”

81

notsneaky 12.08.15 at 10:45 pm

Not quite. If steady state then r constant. But r constant does not automatically imply steady state. You’re denying the antecedent.

82

MisterMr 12.08.15 at 11:46 pm

Piketty is measuring wealth, not productive capital in the usual sense of the term. So if the expected rate of return on wealth is fixed by, for example, psychological reasons, then the amount of wealth will simply grow to match r, and if the share of profits grows, wealth grows relative to total income.

For example, suppose a completely agricultural economy where total income is 100 apples a year, and the sharecroppers get 80% of the product. If the landlords have an expected rate of return of 4%, they will buy the land until the price of land is pushed up to 20X100/4=500 apples. If then for whatever reason the sharecropper share falls to 70%, the price of land will shot up to 30X100/4=750 apples. This regardless of population growth, productivity growth, or even inflation, since we are measuring the value of land vs. the value of total income, not material quantities of stuff.
The independent variable is the wage share, not the quantity of “capital” (actually wealth).

This is a shorter myself at 57.

83

T 12.09.15 at 12:14 am

There are a series of logical and data issues surrounding the claim that r greater than g. Worth taking a look before weighing in on the meaning and technical merits of CIT21C. You can google Rognlie and find various high and low quality descriptions of the papers as well. But here are the originals and a discussion by DeLong.
http://www.mit.edu/~mrognlie/piketty_diminishing_returns.pdf
http://www.brookings.edu/~/media/projects/bpea/spring-2015/2015a_rognlie.pdf
and
http://delong.typepad.com/sdj/2014/06/over-at-equitable-growth-daily-piketty-matt-rognlie-has-a-first-rate-critique-thursday-focus-for-june-12-2014.html

84

Peter T 12.09.15 at 2:39 am

I agree with notsneaky that Piketty’s label of r>g as a “law” is misplaced and irritating.

Another mistake Piketty makes is to use the term “capital” for wealth. His argument only makes sense if r is the rate of return on wealth and wealth is any income-producing stock (not necessarily an addition to production). A near-constant r is then a function of a broad class of political economies, not a factor of production. A lot of the critiques referenced in posts assume that r is the rate of interest or return on capital investment, and so subject to diminishing returns.

Further, notsneaky is right. r>g does not in itself imply an ever-larger share going to wealth-owners. Whether it does depends on how much is re-invested, and who gets the g. Again, the models in a number of critiques assume a single-class society, with individuals differentiated in capabilities and assets. Piketty only makes sense if you assume a hierarchical order – a minimum of two classes.

Before I yield entirely to notsneaky, I’ll note that “random shocks” is another term for “living”.

85

Rakesh Bhandari 12.09.15 at 3:08 am

Are people here seriously arguing that after-tax r may approximate g? I am not economist, but I don’t think we need to get into the intricacies of the Cobb-Douglas production function and neo-classical growth theory to grasp what is going on–see pp. 215ff of Piketty’s book . That r may not be greater than g is not the claim DeLong is considering–that involves the elasticity of substitution and whether any increase in Piketty’s β is at least counteracted by a fall in r so that Piketty’s α does not increase. But note again that Piketty thinks his β will increase due to reversion of growth to historical averages–this is key to his argument– so the increased volume of capital will counteract any possible declining r, resulting in his α trending upward.
Again I think Piketty would have been on stronger ground to abandon any theory based on the marginal productivity of capital and find theoretical support in the Cambridge Capital critique that he unpersuasively critiques.

86

Rakesh Bhandari 12.09.15 at 3:56 am

If we don’t accept Piketty’s explanations, then how do we want to account for the the movement of his α and β over the last two centuries? He certainly seems not only to have demonstrated their movements through careful and heroic data gathering but also given a coherent explanation for them. What are the grounds for rejecting his careful reasoning that both α and β will likely rise (along with the percentage of wealth that has been inherited) in at least the medium term?

87

notsneaky 12.09.15 at 4:38 am

Rakesh, I’m assuming you’re referring to those papers/critiques by Rognlie. That’s really a different argument then the one we’re having here, but yes, Rognlie is right.

If your production function is Cobb-Douglas and you have a fixed saving rate then whether r is greater than g depends on whether that saving rate is greater or less than the share of capital in income. However, whatever the case, that capital share is constant and r vs. g is unrelated to that kind of inequality.

If your production function is Cobb-Douglas and you have an endogenously varying saving rate then r is always greater than g (if consumption is increasing along with income, i.e. if you’re at or “below” the steady state). However, whatever the case, that capital share is constant and r vs. g is unrelated to that kind of inequality.

So you need a non-Cobb-Douglas production function. However, even with that, in steady state r and g will be constants, capital’s share will be constant and the steady state will look very similar to Cobb-Douglas. So instead, Piketty switches tack.

Rather than talking about the steady state (which is EXACTLY what he is assuming when he is discussing his alpha and beta) he waves some hands about the transitional dynamics. If you have a non-Cobb-Douglas production function and an endogenously varying saving rate then yes indeed, you can have that along the transition to steady state r is greater than g and capital’s share increases over time.

But, and this is what Rognlie and others point out (and once you get yourself unconfused by Piketty’s obfuscation, it’s sort of obvious), in order for that to happen you have to assume that capital and labor are substitutes. More substitutable than Cobb-Douglas. In fact, it must be possible to produce output with *only* capital or *only* labor. Piketty claims (though buries the claim) that this is indeed the claim. No one who has ever worked on this or related topics thinks this is true. Whether you estimate this substitution elasticity with macro or micro data, old data or new data, capital and labor are not substitutes. Piketty knows, or at least should know this.

(incidentally the Cambridge folks tend to assume the polar opposite, that capital and labor are perfect complements. Although I’m not sure how exactly Cambridge Capital theory would help Piketty out here))

88

Rakesh Bhandari 12.09.15 at 4:41 am

This is obfuscation–the capital share is not constant. Piketty shows that, so if your Cobb-Douglas function insists that it has to be–for reasons you could better specify–then you should ditch it as having been based on limited empirical information, which is what Piketty underlines.

89

Rakesh Bhandari 12.09.15 at 4:45 am

It may be that the recent rise in α and β cannot be connected to Piketty’s mechanisms but to asset inflation in real estate. But again I think Piketty is getting a bad rap here. He certainly recognizes that his β can rise and fall as a result of bubbles and their bursting, but he also argues that s/g accounts for the long-term K/Y ratio.

90

Rakesh Bhandari 12.09.15 at 4:56 am

See p. 221 of Piketty; and then note http://newleftreview.org/I/151/ian-steedman-robots-and-capitalism-comment
I already cited this.

This is how Piketty could have been helped Cambridge Capital Theory which does not determine shares in terms of marginal productivities but in terms of a political struggle over the distribution of a surplus that is determined by technology. Piketty is essentially arguing both that the available surplus, if anything, may will increase as a result of robotization and that capital as a result of increasing power may well be appropriate a rising share of it, yielding quite possibly a rising r.

91

notsneaky 12.09.15 at 8:05 am

“This is obfuscation–the capital share is not constant. Piketty shows that, so if your Cobb-Douglas function insists that it has to be–for reasons you could better specify–then you should ditch”

The point is that when we “ditch” Cobb-Douglas it actually gets *worse* for Piketty. In that case, if capital and labor are compliments (they’re both necessary for production to take place) then we should observe lower inequality when r-g is high, the opposite of Piketty’s whole thesis. And pretty much any serious study, from any corner of the political spectrum that’s been done comes to the conclusion that capital and labor are indeed compliments in production.

Now, Piketty at this point basically says “well according to my hypothesis inequality should should be when r-g is high, so therefore capital and labor must be substitutes”. In other words, he assumes his conclusion (just in a… obfuscated, way) rather than actually providing evidence for it.

No one’s arguing that capital’s share is constant (well…maybe, but anyway, that’s another argument, not this one) . The argument is that Piketty does not provide an explanation for why capital’s share has risen. At least not with this whole r-g thing.

92

notsneaky 12.09.15 at 8:07 am

@92 But this too misses the point. In the CCC kind of theory the factorial distribution of income is pretty much indeterminate – that’s what “does not determine shares in terms of marginal productivities but in terms of a political struggle over the distribution of a surplus” actually means. But if it’s about political struggle, then it’s not about r vs. g.

93

Bruce Wilder 12.09.15 at 8:23 am

if it’s about political struggle, then it’s not about r vs. g

I am afraid I do not quite see how that must be true.

94

Bruce Wilder 12.09.15 at 8:43 am

Piketty’s r is a return on accumulated claims on income, aka wealth. It could be government debt, it could be land, it could be economic rents associated with a business. There is nothing to suggest that “it must be possible to produce output with *only* capital or *only* labor” enters into it at all. There is nothing about production inherent in the accumulation of wealth.

95

notsneaky 12.09.15 at 9:27 am

@95 Trivial example. Suppose we live in a socialist utopia where workers own the means of production and all have equal shares in the economy’s output. For whatever reason r is greater than g. Is anything going to happen to inequality in this economy? (and if capital ownership is unequally distributed then there’s actually nothing about r greater than g that must lead to greater concentration)

@96 I’m a bit tired of trying to explain it. You can read those links provided by T above in 85. And keep in mind that when you say stuff like “There is nothing about production inherent in the accumulation of wealth.” you’re not arguing with me, you’re arguing with Piketty who’s the one that made the particular choice of this framework.

96

Peter T 12.09.15 at 9:45 am

” Suppose we live in a socialist utopia where workers own the means of production and all have equal shares in the economy’s output”

THEN THERE IS NO r

97

Rakesh Bhandari 12.09.15 at 11:22 am

OK. We have Piketty’s “mysterious” reasoning that if those who enjoy a sharply disproportionate share of wealth are enjoying their wealth grow at a faster rate than the income of those who labor–that is r>g–then the capital/income ratio will tend to rise, the capital share of income will likely rise (though somewhat moderated by a fall in r), the concentration of wealth may well rise (those already with more wealth will be able to add to their wealth faster on account of their r being relatively higher than the average), the share of the wealth and income of the top centile and .1% will likely rise and the share of wealth that is inherited will also likely rise.

All this seems very mysterious to critics of Piketty who claim that there is no obvious connection between his fundamental inequality and inequality and that his having created the data to show that over two centuries r has tended to be greater than g and that inequality has tended towards disturbing and morally incoherent levels–except during times of war, revolution and political chaos–should not be much grounds for concern.

After all, it could be that in the course of accumulation, s, g and r–which are endogenous variables or variables in mutual relationship–could change in such a way that Piketty’s beta and alpha won’t rise as he expects.

No matter that Piketty tries to show that the best ranges that we can give for s, g and r on the basis of historical data painstakingly collected give us every reason to believe that the market left to itself will generate dystopian levels of inequality in the ownership of wealth and income.

Our economist model builders tell us that it is not necessary that things will work out the way Piketty thinks likely–r or s could sharply fall after all in the course of accumulation; g could miraculously reach Jeb-projected levels– and they rejoice at every admission of uncertainty Piketty makes of the future trajectory of market society.

Piketty after all does not claim to foretell the future. Though all Piketty does is let history set the range for the variables in his equations, he is accused of the same historicism with which Popper tagged the Marxists (we’ll leave aside the strangeness of Popper criticizing Marxists for predictions of cataclysm in a work written in exile from the Nazis).

All that said, notsneaky who I think is saying similar things to Acemoglu and thus really is no political bedfellow of Anderson of the OP writes: “if capital and labor are compliments (they’re both necessary for production to take place) then we should observe lower inequality when r-g is high.” Please explain why that would be the case, notsneaky.

98

Rakesh Bhandari 12.09.15 at 11:30 am

@94 misses my point, which was that CCC would give Piketty a theoretical basis to show that the usefulness of technology, plus the power capital gains in the course of robotization, would enable capital to enjoy an r sufficiently greater than g for his alpha to rise. My point was that r likely being greater than g is both the result of politics and technology; in fact Piketty sees it the same way.
Now what he does not understand is that from a Marxist perspective r can remain positive but still not be high enough to yield a great enough mass of surplus value to sustain the accumulation process. That is, Marx’s theory does not predict or depend on r being driven to zero. That it remains positive and greater than zero is perfectly compatible with it having declined to the point of creating a crisis in economic reproduction.

99

TM 12.09.15 at 11:46 am

notsneaky 83: Your *definition* of steady state includes the condition r constant. Therefore it’s a tautology to say (as you did) that every steady state economy has r constant.

75: “I don’t know what “Capital depreciation will end up swallowing the whole of GDP” means.”

1. A portion of production (GDP) goes toward capital replacement (depreciation) and a portion goes toward capital creation (net investment). These amounts are therefore not available for consumption.
2. If r>g and r is interpreted as the rate of growth of capital (or wealth), or a close approximation) (as Piketty does), then capital stock relative to GDP (the capital-income ratio) increases. (*)
3. Increasing capital-income ratio implies that an increasing portion of GDP is used for capital replacement and investment. If this process continued, there would be a point when nothing is left for consumption, which of course is physically impossible, but that is the implication of r>g, and with Piketty’s numbers (3-g about 3.5% over a period of centuries), that point would have been reached quickly. Of course this is NOT a steady state economy.

The argument in 2., which is the heart of Piketty’s r>g argument, makes superficial sense, but it really makes no sense when looked at closely. Piketty’s own empirical data (http://piketty.pse.ens.fr/files/capitalisback/F3) say that capital-income ratio was constant at 700% from 1700 to 1900, and has been steadyly increasing (from a much lower level) since 1950. But his r-g data (http://www.nybooks.com/articles/2014/05/08/thomas-piketty-new-gilded-age/ figure 1) are totally inconsistent with that picture – he says r was always much greater than g *except* in the 20th century, including the period sinced 1950. So empirically, nothing makes sense.

(*) Here is how Krugman explains Piketty, from the NYRB review:
“At the time [19th century], owners of capital could expect to earn 4–5 percent on their investments, with minimal taxation; meanwhile economic growth was only around one percent. So wealthy individuals could easily reinvest enough of their income to ensure that their wealth and hence their incomes were growing faster than the economy, reinforcing their economic dominance, even while skimming enough off to live lives of great luxury.” Inequality increases because the capital-income ratio increases, because capital grows at a faster rate than GDP (r>g). But again, Piketty says empirically the capital-income ratio stayed constant despite r>g (and it increased after 1950 despite r<g).

100

Lee A. Arnold 12.09.15 at 11:49 am

Notsneaky #80: “…even by the standards of social science “laws” (say Gresham/Copernicus’ Law or the Iron Law of Oligarchy) it doesn’t do too well.”

Hold on, Gresham’s Law is essentially a tautology (“that which won’t have greater value, won’t have greater value”). But the Iron Law of Oligarchy is an empirical result that needn’t always hold true, similar to r>g. What’s the difference?

101

Robert 12.09.15 at 1:43 pm

“Trivial example. Suppose we live in a socialist utopia…”

Well, my neoclassical example (see paper linked to by my name) is of pension capitalism, where retired workers live off their assets accumulating interest. And I have an asset that can yield a real rate of return (because of price changes), but does not function as an input into production.

I do not say this model is non-trivial, albeit I lay out the math in too much detail in appendices. I had fun figuring this out, especially stability.

Anyways, it is clear something more than r > g needs to hold for inequality to increase.

102

Trader Joe 12.09.15 at 3:43 pm

Does Picketty address the role of leverage/debt?

Asset/wealth holders can use leverage to magnify returns such that even when r falls short of g, they can still compound wealth. Those lacking assets not only cannot use leverage to compound wealth, they more commonly use debt to substitute for spending which all but guarantees g g alone.

I’m not a student of Piketty, so maybe I’m completely off base, it just seems intuitive. Rich people I know borrow to invest, poor people borrow to live.

103

Bruce Wilder 12.09.15 at 3:50 pm

notsneaky And keep in mind that when you say stuff like “There is nothing about production inherent in the accumulation of wealth.” you’re not arguing with me, you’re arguing with Piketty who’s the one that made the particular choice of this framework.

I think Piketty did a supremely clever thing by confusing wealth and capital, but I am not confused. Without the confusion of wealth and capital, he cannot get a hearing from the mainstream economists, whose thinking is framed by Solow-Swan growth models. The struggle of economists over r > g is a struggle to come to grips with the shortcomings of a model of growth that features an aggregate production function and an accumulating capital, but does not distinguish wages from rents. Not Piketty’s model, so much as their own. Piketty does not have the answer; he has some facts, and he has cast those facts in a way that challenges the orthodoxy, by confusing them and drawing them into furious calculation and critique.

I appreciate that one of the shortcomings of Piketty’s exposition is the oversimplification he chose to make in not distinguishing rates of depreciation and reinvestment (the reproduction of capital) in his analysis. This oversimplification has consequences. That is what I get from T’s links @ 85.

Wealth and capital are NOT the same thing. Like I said, Piketty confounds the terms and concepts. His bad, I guess. But, in doing so, it seems to me he opened up the discussion to the political determinants of the distribution of income. In the full scope of Piketty’s vision, it is possible to see that private Wealth invests in rent-seeking; any investment in productive capital is purely incidental. The accumulation of private wealth can be a bad thing for society and the inexorable accumulation of wealth is a political force affecting the distribution of income. I would not want this obscured.

104

reason 12.09.15 at 3:58 pm

TM
it seems to me that your concept of capital – and that that Piketty is using are different – r includes rent (including rent from land, IP and outstanding debts) and is not a straight return from economically productive capital.

105

reason 12.09.15 at 4:00 pm

Bruce @105
Sorry, I see I have made pretty much the same point as you did. My answer referred to TM @101 .

I’m not sure why this is so confusing for some people.

106

notsneaky 12.09.15 at 5:01 pm

@98 Huh? Why not?

107

TM 12.09.15 at 5:08 pm

reason 106, part of the difficulty is that Piketty’s use of terms like capital and income is not consistent. Nevertheless, I don’t see how the problems I point out at 101 go away by tweaking the definitions. After all, I am referring to Piketty’s own data concerning r, g and the capital-income ratio. These data are simply inconsistent with Piketty’s own interpretation. So please don’t tell me I should use a different definition. It shouldn’t matter.

108

Bruce Wilder 12.09.15 at 5:22 pm

TM @ 109

Piketty’s “Second Law” says that the ratio of accumulated Capital to total Income approaches an an asymptote or limit, the ratio of savings to growth. Did you miss that part? Help me out here.

109

notsneaky 12.09.15 at 5:43 pm

“notsneaky 83: Your *definition* of steady state includes the condition r constant. Therefore it’s a tautology to say (as you did) that every steady state economy has r constant.”

A steady state of an economy is defined as a situation where output and capital (and population) grow at constant rate. If output and capital grow at a constant rate then r will be constant. If r is constant however that does not necessarily mean that the economy is or has a steady state (for example open economies with different time preferences and no adjustment costs to capital flows). So it’s not a tautology.

But if you like you can insist on it being a tautology. The point then would be that Piketty wasn’t the first to “discover” that economies can have steady states (i.e. constant r)

110

TM 12.09.15 at 5:44 pm

BW, his data don’t show that behavior. In any case, why do you think this “law” is relevant for my argument?

Btw here’s a source that makes the same point as I did above:
“the assumption that the capital stock is always growing (because net saving is positive) implies that more and more output must be diverted away from consumption towards investment. Eventually, because capital needs to keep rising, it is necessary to devote 100% of GDP to capital formation!”
http://www.voxeu.org/article/piketty-s-second-law-capitalism-vs-standard-macro-theory

111

notsneaky 12.09.15 at 5:50 pm

Rakesh, I’m sorry but you’ve gone off and started roughing up various strawmen, so I can’t really address that.

One more time. Piketty tries to stay firmly in the neoclassical framework. There’s absolutely nothing in the neoclassical framework which would link r-g to inequality. (There’s actually nothing in any framework I’m aware of that would link r-g to inequality but people are getting hung up on criticizing the neoclassical framework while defending Piketty – you can’t have your capital and eat it too ) He waves his hands. And confuses the issue. And uses some accounting identities to rewrite the same thing in different way and calls that a “law”

112

A H 12.09.15 at 5:52 pm

“Further, notsneaky is right. r>g does not in itself imply an ever-larger share going to wealth-owners. Whether it does depends on how much is re-invested, and who gets the g.”

Of course this is true, but it is no criticism of Piketty. The “r > g” framing is not the best, but the relationship of r and g and inequality are strong and obvious with a few reasonable assumptions. Strong enough to call a law I would say.

If you take an economy with a certain saving rate, r and g and hold these constant, the relationship of these factors will lead to a steady state level capital to income ratio. Piketty then give evidence that 1. g can be expected to fall in the future, 2. capital income and wealth are becoming increasingly unequally distributed, 3. r has been relatively constant in the past despite different economic conditions. Take the accounting and stylized facts together and you have much greater inequality in the future.

Much of the criticism from mainstream economists is about the theory chapter in the book. But that chapter was completely muddled and orthogonal to the main thrust of the book that I outlined above.

113

notsneaky 12.09.15 at 6:17 pm

“2. If r>g and r is interpreted as the rate of growth of capital (or wealth), or a close approximation) (as Piketty does), then capital stock relative to GDP (the capital-income ratio) increases. (*)”

But r is not the growth of capital, it is the return on capital.

The growth rate of capital is (saving rate * output-capital ratio) – depreciation rate. Piketty actually does some redefinition here which cause a lot of confusion and obfuscates a lot. He rewrites the growth rate of capital as (output-capital ratio) *(saving rate – deprecation rate* capital-output ratio) and then calls that second term, (saving rate – deprecation rate* inverse of capital-output ratio), a … saving rate. So then you have growth rate of capital = (output-capital ratio)*(“Piketty’s saving rate”)

In steady state then (again, this is Piketty’s argument) this growth rate of capital should equal g (not r!!!!!). So we have g=(output-capital ratio)*(Piketty’s saving rate). Inverting we have capital-output ratio =(Piketty’s saving rate)/g

He then says “Look! If g is zero then capital output ratio will go to infinity!” which is what I understand you’re referring to when you say that depreciation will swallow up GDP.

But this isn’t true. The sleight of hand is hidden in the way that Piketty rewrote the saving rate.

Piketty’s saving rate = actual saving rate – depreciation rate * (capital-output ratio)

So note that if capital output ratio gets sufficiently large then actually Piketty’s saving rate will become zero. So his steady state condition is really zero/zero and there’s nothing here that will make capital output ratio go to infinity (it will actually be a constant)

It’s like one of those tricks that math students like to play on humanities folks where they prove that 1=2 by slyly dividing by zero.

(in Math:

dK=sY-pK

dK/K=s(Y/K)-p <– growth of capital, s is fraction of output reinvested, p is depreciation rate

dK/K=(Y/K)*(s-p*(K/Y))

dK/K=(Y/K)*s' 0 so that K/Y increases over time. Eventually we get to a point where K/Y=s/p. At that point s’=0. And at that point dK/K=0. So both Y and K remain constant and that’s your K/Y right there, a finite value)

(This was one of the instances in the book were Piketty genuinely started to piss me off)

114

notsneaky 12.09.15 at 6:22 pm

Ugh, stupid mark up hates inequality signs.

in Math:

dK=sY-pK

dK/K=s(Y/K)-p this is growth of capital, s is fraction of output reinvested, p is depreciation rate

dK/K=(Y/K)*(s-p*(K/Y))

dK/K=(Y/K)*s’ where s’=s-p*(K/Y) is Piketty’s saving rate

Now in steady state (again, this is Piketty) dK/K=dY/Y=g, so

g=(Y/K)*s’

K/Y=s’/g

Question: if g=0 does that mean that K/Y goes to infinity? No. Suppose initially dY/Y=0 but dK/K is positive. So K/Y is increasing. Eventually we get to a point where K/Y=s/p. At that point s’=0. And at that point dK/K=0. So both Y and K remain constant and that’s your K/Y right there, a finite value)

115

notsneaky 12.09.15 at 6:24 pm

Oh, and Robert, I like your example.

116

A H 12.09.15 at 6:32 pm

Where does Piketty say that savings goes to infinity?

Given r, g and saving behavior you have a steady state capital/output level. That’s all the Piketty is saying.

117

notsneaky 12.09.15 at 7:10 pm

Not saving. K/Y

118

notsneaky 12.09.15 at 7:14 pm

“If you take an economy with a certain saving rate, r and g and hold these constant, the relationship of these factors will lead to a steady state level capital to income ratio. Piketty then give evidence that 1. g can be expected to fall in the future, 2. capital income and wealth are becoming increasingly unequally distributed, 3. r has been relatively constant in the past despite different economic conditions. Take the accounting and stylized facts together and you have much greater inequality in the future.”

First mistake is in “hold these constant”. Second mistake is in identifying capital to income ratio with “inequality”.

119

A H 12.09.15 at 7:31 pm

1. Well if you don’t hold them constant, you need to explain how they change. That’s why there is a long discussion in the book on the historical trends in r and g. Piketty makes a good case that g will fall and r will remain largely the same in the medium to long term, meaning the ratio will increase. He makes a mistake in connecting this discussion with neoclassical production theory, which is just a distraction in my opinion.

2. The capital income ratio is NOT identified with inequality. Instead a large section of Capital is taken up with a discussion of inequality of income from capital, inequality of income in general and the inequality of wealth. If the inequality of wealth and income from capital are increasing at the same time as the capital to income ratio is increasing, you get a rentier society.

120

notsneaky 12.09.15 at 7:42 pm

1. No, on several fronts. Both saving and r are essentially functions of g. Piketty knows this but pretends he doesn’t because explaining how saving and r are related to g would undermine the whole r-g thing. He also pulls another trick in discussing the “case that g will fall”. The g in there is the *long run* growth rate – it has to be because Piketty bases it on a steady state (i.e. long run) relationship. But when he makes the “case that g will fall” he starts talking about *transitional* dynamics – what happens away from steady state. This is inconsistent and incorrect. Second, neoclassical production theory is not just a distraction it’s THE framework that Piketty is working within.

2. Ok, if capital income ratio is not identified with inequality, why do we care? In your post at 114 you (and Piketty does this too) slipped from talking about determinants of capital income ratio to talking about inequality, so a reader can understandably infer that you mean that capital income ratio is connected to inequality (it’s not).

Btw, you said “Take the accounting and stylized facts together and you have much greater inequality in the future.” Your second “stylized fact” is pretty much circular, it assumes the conclusion.

121

TM 12.09.15 at 7:55 pm

115: “But r is not the growth of capital, it is the return on capital.”

The growth rate of capital is strictly speaking less than r because capitalists consume some of the profit rather than reinvesting it. But for Piketty’s argument, only the reinvested part is relevant and he assumes that only a small part is consumed. At least that is the only way I know how to understand the argument, and this is also how for example Krugman explicitly understands the argument, see 101. Inequality is increased not because rich people consume a lot but because they accumulate more wealth. If on the other hand capitalists consume so much that the growth of capital is reduced to g, then the capital-income ratio will indeed be constant and the distribution between labor and capital won’t change. A large difference r-g then means that capitalists get to consume an obscenely large share of income but that share remains constant and no implications follow for inequality.

Now btw your claim that under steady state, r must be constant, makes no sense any more, because there is no reason why the consumed fraction of capital income should be constant. Unless, of course, that too is presupposed. Forgive me but that kind of “economic law” – “if we make a lot of unrealistic assumptions, then some banal consequence can be deducted from accounting identities” – doesn’t impress me too much.

122

A H 12.09.15 at 8:01 pm

1. Piketty shows empirical evidence that r is relatively constant. Is this evidence wrong?

Of course the transitional dynamics are what is important to Piketty, since a central thesis in Capital is that the world wars caused the western economies to move away from long term steady state. What matters for the dynamics in capital to take effect is not the long run growth rate of neoclassical econ, but the average growth rate. It’s the average rates that cause the stock – flow steady state to be reached over a period of time (NOT the same as long term equilibrium in neoclassical models).

There is really only one chapter in Capital that deals with neoclassical theory, and you could cut it out and the book would be the same, (better even). But I agree that Piketty has been at defending his work from within the neoclassical point of view.

2. Because if the capital income ratio is increasing and wealth inequality increasing than the power of the wealthy is increasing at the same time as concentration of wealth. There are unexplored political implications of this as the OP points out.

123

Jurriaan Bendien 12.09.15 at 8:27 pm

A simple way to understand Piketty:

1) Capital accumulation can occur in two basic ways: either through an increase in the total net stock of wealth (new valued added), or else by an amassment of wealth which is transferred, at the expense of other people. Usually both these modes of accumulation occur at the same time in varying admixtures, but one usually prevails over the other. In the boom, everyone can make gains, even if the gains are unequal, while in a serious slump, the gains of some are mainly at the expense of others.

2) The more equity you own, the more money you can make, at the very least because you can invest it at a higher interest rate, but also because you can borrow more money and use it to extract more profit. The effect is, that the concentration of wealth tends to accelerate (other things remaining equal), meaning that the more capital you own, the faster it will grow too.

3) Piketty assumes a very broad definition of capital, which includes the residential housing stock that grew enormously in the last half century or so. Consequently, his profit rate concept assumes the inclusion of the housing stock, and the income from housing. If however the housing stock and income from housing is excluded from the figures, then his equations do not work anymore.

124

Bruce Wilder 12.09.15 at 9:06 pm

TM @ 112 Per Krusell and Tony Smith are hacks doing a hack job, OK?

This: “the assumption that the capital stock is always growing (because net saving is positive) implies that more and more output must be diverted away from consumption towards investment. Eventually, because capital needs to keep rising, it is necessary to devote 100% of GDP to capital formation!” is clearly idiotic. It is reducing Piketty to an absurdity, by deliberating misunderstanding what Piketty said. Piketty is not without fault in this, but it is possible to understand him, despite the fault.

I told you how Piketty formulated his “Second Law” @ 110: “the ratio of accumulated Capital to total Income approaches an an asymptote or limit, the ratio of savings to growth.” Any reasonable values for savings (s) and growth (g) will result in ratios of Capital to Income that may be worrisome socially, but are finite and within historic bounds. As notsneaky points out, the formula is not correct — it is a serious oversimplification; it leaves out a term for depreciation and the necessary distinction between net savings and gross savings. As the capital stock grows, you’d expect more and more of gross savings to go to simply reproducing depreciating capital. That makes an explosion in the Capital to Income ratio as growth falls to zero an erroneous extrapolation.

The trick that Per Krusell and Tony Smith are playing is to overlook the subtlety introduced by Piketty confounding Capital as a-Factor-of-Production-that-may-accumulate (the neoclassical concept) with Wealth in general, wealth in general simply being financial/legal claims on income. The neoclassical concept makes savings into a diversion of resources from consumption to investment, aka “capital formation”, aka additions to the capital stock.

Piketty says of his formula:

“the law β = s / g is valid only if one focuses on those forms of capital that human beings can accumulate. If a significant fraction of national capital consists of pure natural resources (i.e., natural resources whose value is independent of any human improvement and any past investment), then β can be quite high without any contribution from savings. . . . the law β = s / g is valid only if asset prices evolve on average in the same way as consumer prices. If the price of real estate or stocks rises faster than other prices, then the ratio β between the market value of national capital and the annual flow of national income can again be quite high without the addition of any new savings.”

Piketty’s concept of “Capital” as wealth (claims on income), does not require the diversion of savings from consumption into capital formation in order for the ratio of Capital to Income or the share of Capital’s claims in Income to rise to troubling heights.

I have mixed feeling about this aspect of Piketty’s work. As a matter of measurement, I think it is defensible, while the neoclassical construct of capital as an accumulating factor of production ought to be more seriously questioned. Piketty does not do much of this questioning, but he exposes the vulnerability.

On the other hand, there are many times when it seems to me that he needs to inquire into this distinction between financial claims and synthetic factor resources more closely, and to sustain it in his narrative, and he fails to do so.

125

Rakesh Bhandari 12.09.15 at 9:13 pm

@122. notsneaky writes: ‘But when he makes the “case that g will fall” he starts talking about *transitional* dynamics – what happens away from steady state.’

So Piketty is interested in what you guys may call the traverse or transitional dynamics–what exactly the problem here?. We are moving from one steady state to another as a result primarily of g likely falling from 3.0 to 1.5 percent as a result of the ending of the post-war boom, global catching up and demographic slow-down (so I won’t entertain your question of what would happen were g to fall to zero as this isn’t relevant to the projection).

This gives us a new beta and thus a new K/Y, perhaps climbing to 7.0 (10%s/1.5% g). Piketty would characterize this as patrimonial capitalism. Once that new K/Y is reached, it can enter an economic steady state, but Piketty thinks the problem will be sociological: we will have re-created patrimonial capitalism that is normatively out-of-synch with widely shared meritocratic values. He expects a social reaction or what Boltanski and Chiapello would call a sharpened social critique of capitalism.

Now perhaps you are arguing that as g stabilizes at a lower level, we should expect changes in s and r. Yes, they will doubtless change but all the data that Piketty provides gives us no reason to believe that they will change in such a way to counteract the uncanny return of a rentier society.

126

Rakesh Bhandari 12.09.15 at 9:24 pm

This criticism of Piketty ignoring depreciation do not make sense to me. He uses GDP data that have already accounted for depreciation. This is where he is getting his s and g from. Perhaps in some of the econometric work on the elasticity of substitution there is some confusion about gross and net saving, but I don’t see how that affects the basic data that he uses.

127

Rakesh Bhandari 12.09.15 at 9:44 pm

Larry Summers has already made a lot of these criticisms–Piketty simply assumes that the returns will be fully invested (wrong) and that he has set the elasticity of substitution too high due in part to his confusing gross and net savings. I really don’t find these criticisms compelling, but Summers does raise one concern that I share.

Even if Piketty is right about a slow transition from what he would characterize as a petit rentier to a rentier society in which the capital share of income rises to levels not seen since the end of the long 19th century and capital income becomes increasingly concentrated among the 1, .1 and .o1 percent of the population whose wealth comes to be mostly inherited–even if society evolves into such a disfigured state–there will be likely be other problems along the way that may destabilize society–high unemployment in protracted slumps, job displacement due to new technologies, popular anger at the bailout machinery, fascist scapegoating of immigrants, industrial conflict and of course ecological disaster.

Piketty himself raises the problem of the clash of temporalities or different time scales (p. 286ff). The return of rentier society is a long-term evolution, but there may be a lot of action in the short and medium term, and the prospect of a long-term evolution into a rentier society may have little salience or resonance in the face of more immediate problems.

128

notsneaky 12.09.15 at 10:00 pm

“The growth rate of capital is strictly speaking less than r because capitalists consume some of the profit rather than reinvesting it”

No. I mean, it could be, but it could not be. I have no idea where you got this. I think you’re confusing some concepts here.

129

notsneaky 12.09.15 at 10:08 pm

Suppose one unit of capital produces two units of output. So Y/K=2. Out of those two units capital owner gets one unit and worker gets one unit. Capital owner saves 3/4 of their income. This turns into 3/4 units of new capital. Existing capital depreciates at 10%.

r=(Y-K)/K=1
Growth rate of capital = (3/4)*2-.1=(3/2)-.1 which is greater than 1

Specific units or assumptions don’t matter here. Just a simple way of illustrating that is not true that “The growth rate of capital is strictly speaking less than

130

notsneaky 12.09.15 at 10:13 pm

“1. Piketty shows empirical evidence that r is relatively constant. Is this evidence wrong?”

No, just irrelevant. If r-g does not matter for inequality then it doesn’t matter what is happening to r. Just like if Y/K doesn’t matter for inequality then it doesn’t matter what is happening to Y/K.

Of course the transitional dynamics are what is important to Piketty

“Of course the transitional dynamics are what is important to Piketty”

This is also wrong but I understand your source of confusion since Piketty himself is very very confusing here (to a point where it almost seems to be on purpose). You can’t derive a relationship based on long run steady state theory and then do a slight of hand and justify your theory based on what happens out of steady state.

Y/K=s’/g is a long run relationship. It says absolutely nothing about what happens during transition to steady state. You can analyze what happens during transition based on Y/K=s’/g

131

notsneaky 12.09.15 at 10:14 pm

“Capital accumulation can occur in two basic ways: either through an increase in the total net stock of wealth (new valued added), or ***else by an amassment of wealth which is transferred, at the expense of other people.*** “

Not in the aggregate.

132

notsneaky 12.09.15 at 10:29 pm

“So Piketty is interested in what you guys may call the traverse or transitional dynamics–what exactly the problem here?”

Because his “theory” is based on a steady state condition. The “g” in that formula is the long run g. It is not the g that occurs during transition. You can’t defend your theory of the long run by discussing what may or may not be happening during transition.

133

Rakesh Bhandari 12.09.15 at 10:35 pm

Yes, I agree, and I also agree that P should not have said s and g are independent (Steven Pressman criticizes but ultimately saves Piketty on this by arguing that s will not decline as much as g, so Y/K will rise with the decline in g). What I meant to emphasize (and I think P believes) is that during the transition or traverse there likely will be destabilizing social critique.

134

dsquared 12.10.15 at 1:32 am

“(or in other words, if you don’t like this approach, talk to Piketty)”

I have done (or rather, because I am both lazy and unimportant, I’ve let Jamie Galbraith do so and then said “yeah”). Rakesh is right – the fact that Piketty is really ignorant about capital theory is a major weakness in the book, but less important in the scheme of things than the fact that the world appears to be on a worsening path and that a big part of that path involves rates of return on investment which are higher than the growth rate

135

Peter T 12.10.15 at 2:06 am

notsneaky @ 108: @98 Huh? Why not?

Because r, as Piketty explains (but then confuses by using the term “capital”) is the rate of return on wealth. Wealth is the present value of claims on income from ownership. No ownership of wealth, no r.

A point about r made at comments 17, 29, 30, 35, 44, 46, 84, 86, 96, 105, 106 and 126.

Think about the difference between “capital” as used in Jane Austen and Balzac and in economics texts.

136

notsneaky 12.10.15 at 4:16 am

@137 Presumably a socialist utopia would still have capital, and it would still have wealth, and this capital would earn a return, and this wealth could be invested and would also earn a return. So no, it would still have a (possibly negative) r.

And you do realize the book is called CAPITAL in the 21st century, not WEALTH in the 21st century?

137

notsneaky 12.10.15 at 4:20 am

@dsquared, well we’re not talking about the grand scheme of things, we’re talking about Piketty’s book. You might wish Piketty had written some other book but that’s a different conversation.

All I’m saying here is that Piketty’s whole “high r-g causes inequality” thing is nonsense within the framework he has chosen (and probably within any imaginable framework). And that is the major thesis/original contribution of that book.

138

Peter T 12.10.15 at 4:45 am

@138

A socialist utopia would have capital, and invest and earn a return. It would not have wealth, because no part of the capital would be separately owned.

Socialist utopia: everyone collectively owns an orchard. All eat apples; some apples are set aside and used to extend the orchard (so capital equals orchard plus seed apples). There is no wealth.

Non-socialism: some people own parts of the orchard and take rent in apples. They eat some and save the rest to extend their holding, either by planting or by trading. Wealth is the value of their rents.

IIRC Piketty explicitly defines “capital” as wealth in the sense above. He is using the term as say Jane Austen would have used it: to refer to income-producing assets regardless of whether the income results from productive assets, rents, taxation or capital gains (‘her fortune consisted of no more than 5,000 pounds in consols”).

139

notsneaky 12.10.15 at 5:01 am

I don’t think Piketty’s Jane Austen literary examples are Piketty’s definition of capital. They are just examples, which are suppose to illustrate something or other. Whenever Piketty starts writing down equation, including his “laws”, he’s squarely in the neoclassical definition of capital. True, he switches the definition back and forth (partly because at the end of the day you have to somehow measure “wealth” and return on capital so you got to make a choice of what is included and what is excluded), which just adds to the confusion.

But this is really semantics. You don’t like the socialist utopia example? Fine. Think instead of a capitalist utopia where the claims on capital (orchard) are equally distributed. Then here also it doesn’t matter what happens to r vs. g. You don’t like that example? Look at Robert’s example. You don’t like that one? Look at the standard growth model. There really is no framework out there which will give you that r-g affects inequality except for that one discussed in the JEP paper I linked above and there too there’s some implausible assumptions going on.

140

TM 12.10.15 at 8:39 am

125: “The more equity you own, the more money you can make, at the very least because you can invest it at a higher interest rate, but also because you can borrow more money and use it to extract more profit. The effect is, that the concentration of wealth tends to accelerate (other things remaining equal), meaning that the more capital you own, the faster it will grow too.”

That argument is pretty trvial, and it is independent of r>g. Notice, once again, that this process worked extremely well post 1950 despite r<g.

126 BW. “This (…) is clearly idiotic. It is reducing Piketty to an absurdity, by deliberating misunderstanding what Piketty said. … Any reasonable values for savings (s) and growth (g) will result in ratios of Capital to Income that may be worrisome socially, but are finite and within historic bounds. As notsneaky points out, the formula is not correct — it is a serious oversimplification; it leaves out a term for depreciation and the necessary distinction between net savings and gross savings.” etc.

So it is “idiotic” to take the implications of Piketty’s framework seriously because he arrives at “finite” values by using an incorrect formula and by overlooking the distinction between gross and net savings (bummer). Nice defense. BW, would you at least comment on whether this so-called law is even remotely in line with empirical observation? Piketty’s own data do not show an asymptotic approach to some threshold value.

141

TM 12.10.15 at 9:15 am

131: In your (somewhat far-fetched) example, let’s say K=100, Y=200, r=100% (nice), worker consumes 100, capitalist consumes 25 and reinvests (gross) 75, of which 10 go to depreciation. Net investment then is 65 and growth rate of capital 65%, which *of course* is less than r.

142

ZM 12.10.15 at 10:23 am

If anyone could answer my question above @23, I would still appreciate it.

I don’t understand why r should be equal to or lower than g, unless you don’t want people to make investments.

While very wealthy people holding a lot of capital probably don’t need to make much return on their investments, as a small percentage of a very large figure is still a very large figure, that doesn’t not hold for people who have smaller or small to medium sized investments.

The growth rate in Australia in 2015 was between 2 and 3 percent, the consumer price index rose by between 1.5 and 2.2 percent.

If someone has a small investment and the return on investment is only the rate of growth or lower than the rate of growth, it will be offset by inflation, and their investment is not going to be very worthwhile.

I suppose I don’t mind if the idea is that no one should profit from investments, but that is not why people make investments. So why would you think that the return on investments shouldn’t be higher than the growth rate?

So is Piketty’s argument, that since the rate of return on investments is in the long run higher than the growth rate, there should be a wealth tax for people who hold a lot of capital and don’t need as high a rate of return as small investors?

143

ZM 12.10.15 at 10:24 am

“that doesn’t not hold” unintentional double negative sorry

144

reason 12.10.15 at 10:26 am

ZM
I’m sorry there is some confusion here between real and nominal growth. What exactly do you mean?

145

ZM 12.10.15 at 10:46 am

I am confused about why r should not be greater than g.

If r wasn’t greater than g it wouldn’t be worthwhile investing as you wouldn’t make significant returns. I looked at my Superannuation company and for 5 and 10 year investment streams the annual rate of return is at the lowest (except for one cash option which was much lower) more than 6% p.a. and at the highest almost 11% p.a. Whereas the growth rate this year has been between 2% and 3% which is mostly cancelled out by inflation as in the rises in the consumer price index.

I only have a small Superannuation investment, but if the rate of return was lower than the growth rate, why would anyone want to make investments? (Unless they just wanted to own part of a company and go to shareholder meetings or something like that.)

146

notsneaky 12.10.15 at 3:16 pm

@143 – oops, you’re right re the math. Let me correct it. Each unit of capital produces 5/4 units of output, r=25%. Half the output goes to capital owner, half to the worker. Worker saves zero (that was my unstated assumption above; that workers save too), capital owner saves (3/4) of their income. Capital depreciation is 10%.

So let’s say K=400, Y=500. Capital owner gets 250. Capital owner saves 188. Growth rate of capital is 37% which is greater than r.

It doesn’t matter if the example is far-fetched (what makes it more far-fetched than some other example?) The claim was that r is always going to be greater than growth of capital. The example just shows that this isn’t true.

147

notsneaky 12.10.15 at 3:35 pm

“If anyone could answer my question above @23, I would still appreciate it”

“Would one reason that r is reasonably stable over a considerable time frame…”

Ok, let me try. There’s couple stories you can tell for why the long run r is stable.

One is that each unit of physical capital produces a fixed amount of a consumption good. This means the return on physical investment (building factories, hiring some workers, producing the good, selling it) is constant. By arbitrage, and putting differences in risk aside, then the return on financial investment has to be equal to that constant. If it’s greater then people switch from building factories to buying financial assets. This bids up the price of these assets and lowers their return until the equality is restored. And vice versa. (This is the linear production model story)

Or maybe there’s diminishing returns to physical investment, so that each additional unit of capital produces less and less output. And people are impatient. If you give me 10000$ today, what is the minimum amount that you require in ten years to be willing to make that loan? The percent difference between that minimal amount and the 10000$ is your rate of time preference, a measure of your impatience adjusted for the fact you might be richer 10 years from now. If return on physical investment (building a factory, hiring some workers, producing some output and selling it) is greater than that rate of impatience, you build the factory. Which means more capital. But as the economy gets more and more capital, the most profitable investment opportunities are exhausted and by diminishing returns, the return to building a factory falls. Until it equals that rate of impatience. Arbitrage between physical and financial investment makes r equal to that rate of impatience too. Assuming there’s no long run swings in the rate of impatience, you get a constant r. (This is the neoclassical story)

Or another story would be simply that r is constant because central banks work really hard to keep it that way. (This is essentially New Keynesian story which doesn’t really work here)

148

notsneaky 12.10.15 at 3:39 pm

“I am confused about why r should not be greater than g”

r can be greater than g. r can be less than g. But if r is less than g then the economy is dynamically inefficient – you could cut your saving saving rate, get higher consumption today and also have higher consumption in the future.

All the variables being discussed here are in real terms.

149

Bruce Wilder 12.10.15 at 3:56 pm

K=400, income to capital owner = 250 r = 62%

2nd period
K=400-40+188=548
Y=(5/4)*548=685
g=37%
income to capital owner = 342.5 r= 62.5%
growth in capital 548/500

r=62% > g = 37%

The claim was that r is always going to be greater than growth of capital.

I cannot say I always follow the reasoning that employs these little toy models, but I can do arithmetic and hold a tentative definition of terms in my head for more a moment.

150

notsneaky 12.10.15 at 5:48 pm

Bruce, your 62% is the capital share, not r.

151

notsneaky 12.10.15 at 5:51 pm

Err, sorry, your confusion is confusing me. The 62% is the income-capital ratio.

r is just ((5/4)-1)/1=25%

152

Bruce Wilder 12.10.15 at 7:29 pm

My confusion is confusing me. I did not account for depreciation in calculating the capital owner’s return on invested capital.

r is the return on capital invested, that is, net income from capital

In period one, Capital invested is 400, and the capital owner is getting a gross income from the capital of 250, net of 10% depreciation that’s 210. 210/400 is 52.5%

I don’t know what to call the coefficient 5/4, but it’s not r, a return on capital.

153

Bruce Wilder 12.10.15 at 8:27 pm

The coefficient (5/4) would be the capital to income ratio, I guess, if that makes any sense.

154

notsneaky 12.10.15 at 9:10 pm

5/4 is indeed the capital income ratio. Anyway, the point is that there is no reason why r must be less than or greater than or equal to the growth of capital. It can go either way.

155

Rakesh Bhandari 12.10.15 at 10:56 pm

So what I am getting out of this discussion is that since Piketty has compromised his commitment to neo-classical economic theory by showing that actual historical evidence gives us little reason to accept the elasticity of substitution in the Cobb-Douglas production function, he will not be taken seriously in economics departments and will have the seriousness of historical results and projections better recognized elsewhere in the social sciences. It’s quite possible that the last person who would be surprised by this is Thomas Piketty himself.
Still perhaps the economists will still talk to him since after all he humbly used historical findings such as the rise in his alpha over long periods to criticize neo-classical economic orthodoxy. In this way he is in a much better place than if he had also used Cambridge Capital theory to make the case for why r would remain high enough for a rentier society to emerge; instead he trash talked those heterodox theorists, endearing himself back to the neo-classical economists whom he had just criticized.

156

dsquared 12.11.15 at 12:35 am

All I’m saying here is that Piketty’s whole “high r-g causes inequality” thing is nonsense within the framework he has chosen (and probably within any imaginable framework)

? “The last twenty five years” is an imaginable framework. “Returns on capital being higher than growth means that capitals share of production is increasing” is a perfectly intelligible statement and so is “all the mechanisms which are meant to stop this increase from continuing indefinitely don’t seem to be working very well”.

Which comes back to my original point – economics is the study of the economy, which is a real historical thing that we go out and live in every day. When you say “any imaginable framework”, you’re again talking about abstractions.

I think r > g is a useful law of motion, like the law of motion of government debt. After all, in steady state you can’t have i > g for the nominal yield on sovereign bonds, but it is nonetheless very useful indeed in some situations (like Greece) to do quite a lot of analysis of the drivers of why i > g and why nothing at present looks like it’s able to change that fact.

157

Robert 12.11.15 at 12:52 am

After reading half of Piketty’s JEP response to the symposium on his book, I want to rewrite my critique. It provides a contrast by showing a case in which (1) all three of Piketty’s laws hold and (2) no dynasties inherit wealth and no shocks occur to different households in the same generations. Also, the types of assets that can be owned do not change and are excessively non-diverse, as compared to empirical reality.

158

greg 12.11.15 at 2:05 am

Thank you, Elizabeth Anderson, for your excellent catalog of evils. Thank you also Dan@2 and chris@6 Each of these policies increases r at the expense ofg. Their cumulative activity augments the force and determinism of Piketty’s argument many times. More, that all these mechanisms for upward redistribution are presently active, whereas the further we look to the past the fewer and fewer we find, (into perhaps the 1960’s) suggest that the increase in r tog is accelerating.

However, as the return on ‘capital’ increases, this can only imply that ‘capital’ itself increases, at an ever faster rate. Butg, growing ever more relatively slowly, suggests that real capital is not increasing at the same rate. So, since ‘capital’ represents ownership, the question increasingly becomes, ownership in what?

Consider also, that since g < r suggests that investment in real capital does not return as much as investment in what ever ‘capital’ r represents, investment in real capital is discouraged, compared to investment in- what?- ‘nominal(?)’ ‘capital.’

159

A H 12.11.15 at 4:06 am

Dan Kervick has some very good posts on his blog, Rugged Egalitarianism, that deal with why r > g is important and the issues notsneaky brings up. I tried to post a link but it got stuck in moderation.

From the post “Nothing Magical about Piketty’s Mathematics”:

“…the Second Fundamental Law of Capitalism is neither an identity nor or an approximate identity, but a long-term asymptotic law. We might therefore choose to write it less misleadingly this way:

β → s/g

The import of this law is that, for any fixed national savings rate s and fixed growth rate g sustained over some period of years, β will be moving in the direction of s/g during those years. If β is less than s/g, then the rate of wealth growth will exceed the rate of income growth, and β will increase. If β is greater than s/g, then the rate of wealth growth will be less than the rate of income growth, and β will decrease. Piketty employs this long-term law to explain why some countries which save a lot and grow slowly accumulate large stocks of capital relative to income, and how “small variations in the rate of growth can have very large effects on the capital/income ratio over the long term.” The law thus helps to explain “capital’s comeback” in recent decades…

When combining these theoretical conclusions relating to savings, growth and the capital-to-income ratio with further posits about the rate of return to capital in the description of a hypothetical economy, we need to impose some economically realistic further conditions, even in the case of extremely idealized models, if the posited rates are to describe a possible real-world economy. For example, since the total annual return to capital is rK, and that return can never exceed Y in any economy, then r must always be less than or equal to 1/β. Employing the Second Fundamental Law, another way of putting this is that it is that since 1/β continually approaches g/s, it is not possible to have even a fantasy economy with indefinitely fixed g, fixed s and fixed r, where r exceeds g/s. If these are the initial values, something would have to give eventually.

Nevertheless, for real world economies of the kind for which we have historical experience, rates of return on capital within the historically observed 4% to 5% range can be sustained indefinitely. Indeed, if an economy stabilized with a β of 6, and with relatively stable rates of growth and savings, it could sustain any return on capital less than 16.7%. Even an economy with a β of 10, a value far above the normal range for actually existing economies of the past few centuries, would permit a sustainable return to capital of 10%. Since Piketty observes a consistent 4% to 5% return to capital over this time, there is no tension whatsoever between his observed results and the mathematical model. Nor is there any mathematical reason to doubt that r can, and will, exceed g by a substantial amount in the future, leading to strong forces of income divergence and increasing inequality….

It is also important to note that Piketty’s inequality r > g is not responsible by itself for the strong forces of contemporary income divergence and growing economic inequality. Whether r > g is producing inequality or not depends on dynamics within the society. In a society, for example, in which every adult had an equal ownership stake in the society’s capital stock and capital income, then only differences in returns on labor would contribute to growing or shrinking economic inequality. What Piketty points out is that, given the unequal distributions of capital that actually obtain, r > g is a force for greater wealth and income divergence. And the greater r exceeds g, the stronger these forces will be. There are other factors driving inequality as well, including the rise of “super-managers” and their super-salaries, a phenomenon that Piketty argues is taking place mainly in the English-speaking world. Piketty’s full analysis of the structure of contemporary inequality, and its likely future evolution if the forces driving it are not altered by policy, is not derived mechanically from a couple of simple laws, but is developed over six chapters of the book.”

160

notsneaky 12.11.15 at 6:01 am

“I think r > g is a useful law of motion, like the law of motion of government debt”

It is only useful if it is true and if it is a casual mechanism. I.e. r > g causes greater inequality. Piketty does not successfully explain how this would occur. He tries, sure, but the whole point here, which for some reason you insist on not getting, is that those explanations don’t make much sense. He also does not even convincingly demonstrate that there is a purely empirical correlation between high r-g and increases in inequality.

No explanation (i.e. no framework), no (weak) correlation = nothing.

r greater than g is not a “law of motion”. It is not a “law”. It does not even concern any kind of “motion” (although there are parts in the book where Piketty tries to pretend that it does; one instance that sticks out in my head is where he very sloppily tries to talk about Marx). At best – although this is wrong too – it might describe one kind of steady state versus another (in other words, comparative statics).

The comparison with debt dynamics is a false equivocation because that one actually is a law of motion.

And this *is* important. By emphasizing r vs. g (because he wants to discover a “general law”, which is “groundbreaking” and “sexy”) as some kind of “law” Piketty either purposefully or not, de-emphasizes more plausible, though more mundane explanations. Like “we’ve had 25 years of anti-labor/pro-wealthy policy, but before that we had 25 years of pro-labor/pro-wealthy policy and that’s what caused inequality to increase. Suckers.”

161

notsneaky 12.11.15 at 6:05 am

Obvious typo in that last sentence above; “pro” – > “anti”

162

notsneaky 12.11.15 at 6:06 am

Robert, read the Chuck Jones article in that issue too. He holds back (it’s an academic journal) but it explains the problems pretty well and there are some similarities with your model.

163

Rakesh Bhandari 12.11.15 at 6:26 am

I find this interview with Piketty truly remarkable.
http://www.potemkinreview.com/pikettyinterview.html

164

Bruce Wilder 12.11.15 at 6:53 am

A H

Thank you for that. Kervick has been a thoughtful and intelligent commenter on Piketty. I sought out and read most of the original post.

I had an insight as I read one particular phrase, and I will take a shot at expressing it, though I expect to fail to some extent. Nothing against Kervick or his POV as expressed in the quoted passage is intended.

Here is the phrase: In a society, for example, in which every adult had an equal ownership stake in the society’s capital stock and capital income . . . .

It is inevitable that we should try to understand an idea by trying out counterfactual thought experiments that contradict it. Kervick warns us against wholly unrealistic numbers and turning variables into constants.

We ought to be cautious about ontological errors as well, mistaking fundamentally the nature in being of the phenomena we are speculating about. Imagining for a moment a society exhibiting perfectly equal capital ownership, I started to object in my own mind that Capital ownership is never equally distributed and a louder voice still in my mind, because it would serve no functional purpose in the society. It would be like the hypothetical socialism Peter T offered in a comment above, a society without wealth.

It is easy to forget, in the face of neoclassical insistence that Capital is embodied in artifacts such as houses or tools or brand names that Capital is a social relation: brought into being as promises and debts and functioning socially to organise cooperation, indeed to compel cooperation. It is, I submit, nonsensical to talk of capital in a society of radical equality. Capital is a social institution of an unequal society, of a hierarchical society.

It is also wrong I think to conjure capital without contingency. I know we do it to make algebra sufficient and numerical examples easy, but in its essential being, capital is about social cooperation among unequals, under conditions of risk and uncertainty.

I am not saying anything, I guess, that Peter T did not say earlier, but I wanted to repeat it. Capital is typically a financing of economic activity thru time: it is a lender and a borrower or a creditor or an insurer. It is not the certain flow of income that a discussion of national account categories might suggest, not a property of steady states. It is a bet and an exercise of power in the marshalling of resources.

If we are worried about inequality of wealth, it is because it represents an inequality of not just status, but power. In the cliche, “the rich get richer . . .” there is an awareness that the rich profit from the misfortunes of the poor — quite literally, capital is conduit for those upward transfers. It is against this basic function in the social mechanism of capital that we lay any talk of factors that may act to disperse great fortunes. As power corrupts, wealth extracts.

The tension that Piketty’s identification of Capital with Wealth sets up with the neoclassical concept of capital as an artifactual factor of production is somehow the horizon, where we are drawn into forgetfulness.

I am fading . . . time to sleep . . .

165

dsquared 12.11.15 at 8:13 am

It is only useful if it is true and if it is a casual mechanism. I.e. r > g causes greater inequality. Piketty does not successfully explain how this would occur

This is a quite interesting looking-glass view. You’re making valid points about how “Piketty does not successfully explain how this would occur”, but kind of ignoring the large sections of the book in which Piketty explains how it *does* occur.

If the rate of return on wealth is higher than the growth rate of the economy, then if there’s an initially unequal distribution of wealth it gets more unequal. That is a law of motion, and it’s a locally very important one if conditions are similar to those which do actually obtain at the moment in a lot of important places in the world, including the ones we live in.

166

ZM 12.11.15 at 8:40 am

“I think r > g is a useful law of motion, like the law of motion of government debt. After all, in steady state you can’t have i > g for the nominal yield on sovereign bonds “

I was asking above why r would be lower than g in the long run, since I can’t see why people would make investments otherwise (apart from wanting to go to shareholder meetings).

Most of the growth rate in Australia is cancelled out by inflation as counted by the consumer price index (which I think doesn’t include the price of housing which has grown at a higher rate than the official inflation rate), so your investments would only yield a very small return if r was equal to or lower than g — only around the bank deposit savings rate. And moreover, if r was equal to or lower than g it would mean that investing in general became a riskier endeavour than it is presently since r is the average.

But since you mention a steady state economy — even though investing wouldn’t tied a good return if r was equal or lower than g, I don’t think having r being greater than g would work in a no-growth or de-growth economy.

If the economy was de-growing then businesses would be less profitable or closing, so it would be unlikely that r would be greater than g in that situation.

And if the economy stabilised at zero growth, then r being greater than g would present an even greater problem of concentrating wealth, unless all the profit was taxed so there could be no concentration.

But then you are back to the problem of why people would want to make investments if r was equal to or lower than g (apart from going to shareholder meetings*).

*One of the Melbourne City Councillors (Stephen Mayne) became prominent doing just this, he would buy a small amount of shares, and then go to shareholder meetings specifically to ask thorny questions

167

Peter T 12.11.15 at 9:03 am

I don’t think notsneaky (and many other commenters) can abandon the notion that capital is something real that can be measured (albeit with some difficulty). Wealth – claims on income through title – can be. But wealth is only loosely tied to production.

Capital in the economics texts is a philosophic category, without testable empirical referent. One could, until the mid C19, legally invest in piracy or slaving (and that was capital). One can still invest in enterprises whose returns are based on the destruction of the environment, on the taxing powers of despotic regimes or on various forms of confiscation (and these are capital too). Conversely, many productive assets are unvalued: there are machines working away which have long been depreciated, whose book value is that of scrap. And not just machines – there are bridges, tunnels, canals, landscapes, footpaths and much more (and these are capital in the economic sense, but not accounted). And then there is intellectual, social, institutional capital…

Better to look at what can be measured, what measures play a socio-economic role, how what is counted changes and why.

168

TM 12.11.15 at 9:53 am

148: “Each unit of capital produces 5/4 units of output, r=25%.”
Please think this through. In your example, r=5/4 divided in 2, or 62.5%, which is of course greater than 37%.

169

dax 12.11.15 at 12:25 pm

FWIW (and maybe or probably someone has already made this point), if you want capital to earn less wrt labour, the easiest way is to decrease or limit the supply of labour. An empirical study in an environment where the working-age population is generally increasing exponentially, *should* show that capital is doing better wrt labour. Once population is controlled, labour will do better, and capital worse.

170

TM 12.11.15 at 12:50 pm

161:

β → s/g (where s is the *net* savings rate and β the capital-income ratio) is indeed more or less a law, albeit a trivial one and only meaningful if s and g remain constant in the long run, hardly a reasonable assumption. What it implies is just that if the savings rate exceeds βg, then capital accumulation is faster than g and the capital-income ratio increases. In the steady state, β = s/g and the rate of capital growth (generally k=s/β) is exactly g. (According to Piketty’s data, β was fairly constant from 1700 to 1900 at 7, g rose from .5% in the 18th to 1.5% in the 19th century. Thus the savings rate must have also increased from 3.5% in the 18th to 10.5% in the 19th century).

Notice that r doesn’t figure in this “law” and r>g has no bearing on it. Dynamically, the law predicts that the capital-income ratio increases if the savings rate is large (exceeds βg), and that k in the long run will approach g. In that case, r>g simply implies that the wealthy get to consume a lot. This may be considered undesirable in itself, but that is not Piketty’s argument. Piketty is not concerned with the rich consuming too much, but the rich accumulating too much capital. And for that, r>g is irrelevant – irrelevant according to the model, and also irrelevant according to the empirical data.

171

TM 12.11.15 at 12:54 pm

171: Good point, and worth pointing out that in Piketty’s sense, population stabilization is a bad thing because it depresses economic growth. Piketty thinks that the role of inherited wealth tends to dilute quickly in a fast growing population. There may be a grain of truth in this but trying to solve this problem with population growth rather than with an inheritance tax seems grotesque.

172

TM 12.11.15 at 1:16 pm

Piketty: “Decreased growth – especially demographic growth – is thus responsible for capital’s comeback.”

This is I think very wrong-headed.

173

Robert 12.11.15 at 2:05 pm

If I ever submit my paper somewhere, I am now thinking of using this as an abstract:

This paper presents a neoclassical overlapping generations model in which the rate of growth is positive, income distribution does not become more unequal in a steady state, and the effective real rate of return on savings exceeds the rate of growth. The existence of two assets in the model distinguishes between capital and wealth. In contrast to Thomas Piketty’s empirical results, no dynasties inherit wealth in the model, and no shocks occur to heterogeneous households in the same generation. Without such phenomena, the excess of the rate of return on wealth over the rate of growth need not drive increased income inequality.

174

notsneaky 12.11.15 at 2:13 pm

Re 170, see 153. 5/4 is not r. 5/4 is the income capital ratio.

175

William Timberman 12.11.15 at 2:16 pm

I agree that Piketty’s narrative conflation of wealth and so-called productive capital was less wrong-headed than a number of professional economists seemed to think, especially when backed up with the reams of statistics that apparently made it so difficult to attack him on the grounds of a lack of professional seriousness. Jane Austen? Really? Yes, really.

It was refreshing to see economists forced to chew on something that those of us who are more familiar with the social and political effects of accumulated wealth than with economics have never been able to interest them in. When I look at what’s been happening to the European model of social democracy, and to the tattered remnants of the New Deal’s implied social contract in the US, my first thought has always been wherefore by their fruits ye shall know them. Call it an unbecoming Schadenfreude on my part, but it’s nice to see the pundits and experts confronted finally by an interlocutor far more formidable than I could ever be. The not-so-subtle references to Marx in Piketty’s title, and in his scene-setting, are just icing on the cake.

176

notsneaky 12.11.15 at 2:35 pm

Re 172, why would constant s and g NOT be a reasonable assumption in the long run?

177

TM 12.11.15 at 3:33 pm

notsneaky 176, you are driving me nuts. The capital owner gets half the outout, that is half of 125%, so his rate of return is 62.5%. And re 178, look at the data. g has clearly not been constant. In any case, the question must be turned around: Why on earth should these economic parameters be constant? You cannot assume something to be constant just because you don’t know better. What kind of science is that.

178

notsneaky 12.11.15 at 3:45 pm

Let’s put the first one aside for now.

With g, g is the long run growth rate. Not fluctuations in the growth rate around a trend. Not the growth rate you get as you approach a constant (steady state). The long run growth rate. And it does seem like countries approach a constant g in the long run.

And why should they be constant? Well, for g it would diminishing returns to capital and labor. For s it would be either that the income effect and the substitution effect cancel each other out (maybe unlikely) or exactly because g is constant in the long run.

179

Bruce Wilder 12.11.15 at 3:51 pm

Peter T @ 169

One of the several severe criticisms of the neoclassical use of the production function as a model is precisely that Capital conceived as an aggregate of “real” things — tools, location, reputation, ideas, inventories, organization, etc — cannot be measured, because there is no way to add it all up, without a common numéraire. Once you use money and financial value as a means of measurement, you have given up the concept of a distinct “real” economy — at least you should give it up, because it has no referent.

Capital investment in things, tangible and intangible, is distinctly a phenomena of creating sunk costs. The capitalist entrepreneur, the improving landlord is making irrevocable commitments in concrete, iron, management structure, advertising, because the aspects of the production process that require capital are the irrecoverable activities far in advance of consumption. Timing is everything and the thing about sunk costs is that they cannot matter to subsequent decision-making . . . or bargaining. Neither the actual, divers resources and efforts nor their financial costs ex ante matter to their value as capital ex post. The cost of building a structure or acquiring rights to a location do not affect subsequent negotiations over the rent. The capitalist will calculate in his speculations, of course; choosing what to spend based on expectations about the bargaining that follows. But, when and if a capital investment realizes any return, any income, that return will be in the nature of an economic rent. This is an aspect of capital’s factor income that the concept of the production function can lead analysts to overlook or distort.

One can imagine opportunities to invest are opportunities to do well by doing good in this best of all possible worlds, but the ability to invest with an expectation of return capturable as private income is tied not to resources that may be dedicated to the enterprise, but to an essentially political power to command an economic rent.

The tension in Piketty between neoclassical concepts of an accumulating and productive capital and competing forms of wealth such as government bonds with a more naked basis in political power rests on a stubborn refusal in mainstream economics to recognize even at the first analytic cut as a Krugman might say, income distribution is all politics all the time. Whatever income is realized on a sunk-cost capital investment in production is due to a political institution able to deliver an economic rent. Figuring out how to do this is what MBAs call, finding a viable business model in which to monetize innovation. It is not invisible except sometimes in our imaginations, when we are caught up with the familiar contours of the neoliberal argument.

In the neoclassical argument, money seeks out investment opportunities and as capital accumulates, marginal returns diminish. The accumulation of wealth is self-limiting, capital’s share of income tending toward a limit near a socially optimal structure of production and distribution of income. But, if money is seeking out political power and economic rents, to extract returns, it is not such a happy story and the limits ?

180

Bruce Wilder 12.11.15 at 4:07 pm

In relation to the other thread, on normative justice, my line of thinking does leave me to wonder if capital, like property and coercion, is a practically necessary form of injustice.

181

dsquared 12.11.15 at 4:09 pm

why would constant s and g NOT be a reasonable assumption in the long run?

Never has that Keynes quote seemed more apposite. The data’s in the book, and it covers a period of time long enough that not only are we dead, but over which countries, empires and entire political and economic systems have time to rise and fall. If Piketty is proved wrong in the long term but between times capitalist society has been torn apart by social tensions and ended in bloody revolution, I would consider that a Pyrrhic and moral victory at best, as I dangled from my lamp-post.

182

notsneaky 12.11.15 at 5:42 pm

Piketty is NOT invoking Keynes. If he’s proved right it has nothing to do with Keynes. If he’s proved wrong it has nothing to do with Keynes.

(and actually, empirically a constant long run g and s are quite reasonable assumptions. Kaldor pointed it out long long time ago) You’re going off on red herrings.

183

notsneaky 12.11.15 at 5:45 pm

I’m actually surprised that people are questioning that the long run g can be thought of as constant. Go to google. Type in “real gdp xyz” where xyz is any developed country. Click the little “Images” thingy. Look at the picture.

184

phenomenal cat 12.11.15 at 8:17 pm

Bruce Wilder @181

Your theoretical exposition makes sense to me…or rather I should say the premise that capital investment acts as a (necessary) sunk cost makes sense– I think. But the transition whereby any return or income on the capital investment becomes a matter of economic rent is not clear to me. I can see where this would be the case when capital investment is money, financial instruments, or other such imaginaries explicitly seeking a return. It also clear enough when a given capitalist extracts rents from land or structures. I guess I’m looking for empirical examples of what you are saying.

Say you’ve got a private printing company. Investment includes the land, building, printing machines, computers, and so on. All of these are “sunk costs” necessary for the production of printed material which is sold to customers at some measure above the cost of production. Where, at what point, and how is an economic rent obtained in this process? The investor/owner will surely be looking for a “profit” from the productive capacities of the sunk costs (and labor), but profits, I presume, are not the same as economic rents. What or where are the opportunities for rent extraction in this example?

185

kevin quinn 12.11.15 at 9:52 pm

NotSneaky: I think you are right about r-g and capital share. But what about Jones’ argument connecting increasing r-g, in certain sorts of models, with increasing wealth inequality? Is it too much of a special case, do you think?

186

notsneaky 12.12.15 at 7:08 am

I think in that context Piketty is on much firmer ground and I wish the book had focused more on that rather than the capital’s share and the Y/K ratio. Even there though, once you let r be a function of g some of Piketty’s results disappear. My sense is that this is the right track to pursue but it’s missing something (credit constraints?)

187

TM 12.12.15 at 11:10 am

What is the “long run g” of Germany, Britain, France, China? And, whatever your answer – how do you know? How do you decide what is a short term fluctuation as opposed to the long run trend?

I don’t think this debate is meaningful any more. You wish to discuss Esoterics, I happen to care about empirical reality.

d2, I’ll bite. As shown already many times, Piketty’s r>g theory doesn’t fit the data. What is the matter really? Why is it heresy to point that out?

188

Layman 12.12.15 at 12:13 pm

TM @ 189, that’s a remarkably odd question given the topic of the OP.

189

notsneaky 12.12.15 at 3:21 pm

“What is the “long run g” of …”

About 1.5% for the first three, we don’t know yet for China.

“how do you know?”

Same way we know anything, data and theory.

“How do you decide what is a short term fluctuation as opposed to the long run trend”

There’s some judgement calls there but how you decide mostly has implications for the fluctuations rather than for the long run trend. A non-constant long run g would show up some kind of secular swings (Kondratiev waves and the like) or g getting ever higher or ever lower. It’s pretty clear that for developed economies, that is not what’s happening.

“You wish to discuss Esoterics”

I don’t think the question of whether there is a long run growth rate g and what it is is esoteric. It’s sort of central to Piketty.

“I happen to care about empirical reality.”

90% of time when I see people say stuff like that it turns out to be code for “I’m just gonna make some stuff up out of thin air”

190

dsquared 12.12.15 at 4:26 pm

Even there though, once you let r be a function of g some of Piketty’s results disappear.

“Once you let r be a function of g” just means “once you assert that the phenomena observed by Piketty will cease to obtain in the future” though. If the last fifty years isn’t the “long run”, then we need to worry about short term dynamics too.

191

Walt 12.12.15 at 5:11 pm

90% of the time someone invokes economic theory, it’s because they’ve decided to ignore the data.

Is non-constant g important to Piketty? (I read the book a while ago now, so if it is I don’t remember it at all.)

192

notsneaky 12.12.15 at 5:23 pm

@192 – not really. Shouldn’t the real return be a function of growth? And we’re not talking about a phenomenon here, we’re talking about a theoretical relationship postulated by Piketty. Which doesn’t really make sense.

I don’t understand the last sentence in your comment since it’s not really relevant. And unclear. What is it about the last fifty years that you’re discussing? That inequality increased? Yes, of course we should worry about that but what does that have to do with “short term dynamics”? Or that r was greater than g? Was it? And again, not sure what that has to do with short term dynamics.

@193 – Walt, maybe. But we can’t make sense of the data without theory. And we’re not even talking about theory here – see my suggestion to google images of GDP for developed countries. That’s data right there.

And I think you have it backwards. “non-constant g” doesn’t make much of an appearance in Piketty. That seems to be important to TM, not Piketty. Piketty talks about the long run, constant g. He thinks it’s fallen (i.e. the long run constant level it’s converging to has fallen). But there are parts of the book where he starts talking about short term dynamics in g and capital share and gets it all wrong (both theoretically and empirically)

193

Rakesh Bhandari 12.12.15 at 5:51 pm

I think Charles Jones’ criticism of Piketty has been recommended more than once. I do not have his Macro textbook; does anyone know if he discusses wealth concentration, the distribution of wealth generally, and the capital income share in his textbook? I would guess that he mentions the wonders of compound growth in raising living standards, but is that discussed in relation to the growth and concentration of wealth? Have the Piketty concerns worked their way into textbooks; are his equations discussed along with Solow’s?

194

notsneaky 12.12.15 at 6:22 pm

Jones’ macro book and growth book are fairly standard, which means that no, there’s no discussion of concentration of wealth in it (I am writing this from memory). Piketty’s concerns have not appeared in textbooks and his equations … well, one is just an identity which is not very useful, the other one is the standard capital accumulation equation so it’s not really “his”. It’s been part of any growth model ever since Ricardo or something

195

Rakesh Bhandari 12.12.15 at 6:53 pm

Got it. Thanks.

196

Collin Street 12.12.15 at 7:34 pm

> Shouldn’t the real return be a function of growth?

That’s an empirical question, and it’s Piketty, not you, who’s done the empirical work.

197

notsneaky 12.12.15 at 8:22 pm

@198. “That’s an empirical question” – Not really. Empirical work can only tell you if there is a correlation. Without theory it cannot tell you if r is a function of g or if they’re co-determined by something else.

And no, Piketty did not do THAT empirical work, he did different empirical work (mostly data gathering, which is actually both underrated in general and quite important and very impressive in this case in particular)

198

TM 12.12.15 at 10:09 pm

193: Exactly. Pikettyy first became famous for his empirical work in inequality. Then he wrote r>g and since then, people have been making silly, data-averse and mostly irrelevant claims about long-run trends in the economy, mostly based on the hope that deep knowledge can be deducted from superficially (if at all, see 148, 176 etc.) understood principles. Nobody who understands (let alone has done) real science would ever believe that deep and meaningful predictions about the universe can be distilled from simplified accounting identities.

“Is non-constant g important to Piketty?” Piketty’s “second law”, explained at 172, requires the assumption that g and the savings rate s are constant, or more precisely, that their quotient be constant. When it doesn’t hold, there is no law. Period. Of course, Piketty also points out that historically there have been huge changes in g, and he states that “small variations in the rate of growth can have very large effects on the capital/income ratio in the long run”. If you take that seriously, then it won’t do to look at the chart and say, the trend line is somewhere around 1.5%.

I recognize that simple models relying on simplified assumptions can be useful for certain purposes, but I have no patience for anybody who claims that the behavior of a complex system in the long run can be modeled by assuming certain parameters constant, when there is neither theoretical nor empirical support for such an assumption.

199

notsneaky 12.13.15 at 12:02 am

I don’t think Piketty says that “historically there have been huge changes in g” (and if he does, he’s wrong). He might say that historically there have been huge changes in *the growth rate of output* (and/or capital) but that’s not the same thing.

g = long run growth

observed growth = g + transitional growth

For developed economies “transitional growth” = approx. 0

So there have been … large, changes in observed growth as economies develop. But as economies develop the transitional growth” part goes to zero. And at the end you got just that g. Which is pretty constant.

200

notsneaky 12.13.15 at 12:04 am

” I have no patience for anybody who claims that the behavior of a complex system in the long run can be modeled by assuming certain parameters constant, when there is neither theoretical nor empirical support for such an assumption.” – this is the part where you get wrong. In both respect.

201

ZM 12.13.15 at 12:11 am

I think in the long run g is zero. If you look at Piketty’s graphs he has zero growth until 1700.

There would have been growth when agricultural civilisations began in some parts of the world , but he doesn’t look that far back.

The growth since 1700 would have been based on enclosure, colonisation, science and the industrial revolution etc

There is no reason to think the sorts of growth rates that were common over those 300 years would be long term, it’s more likely zero is the long term growth rate, with periods of higher growth due to paradigm shift style innovations like the development of agriculture and the scientific revolution

202

Collin Street 12.13.15 at 12:55 am

I have no patience for anybody who claims that the behavior of a complex system in the long run can be modeled

You can leave it off there, you know: feedback / differential equation systems of the sort that economists use to model economies have been known to be unstable — “chaotic” — since late victorian times.

Economists are bad at maths.

203

notsneaky 12.13.15 at 2:20 am

@203 That’s Malthus. We’re not there anymore.

@204 And which equations would these be, exactly?

204

Bruce Wilder 12.13.15 at 7:24 pm

phenomenal cat @ 186

Thank you for a question. I’ve essayed several answers, which have all evolved beyond my control into lengthy didactics. The popular use of phrases like “rent-seeking” and “rent extraction” as pejoratives makes feel it is difficult to use “economic rent” in a neutral voice to refer to ubiquitous phenomena and be understood.

To take up your printing company example, the difference between the goal of an accounting profit — that revenue should exceed expenses — from routine operation and a concern about economic rents is the difference between a soldier fighting in a battle and a general planning strategy for the battle and the war. The general makes commitments based on anticipated contingencies, choosing tactical training or lines of supply or positions on the high ground, with an eye to the advantage to be obtained in the fight. Just so, economic rents are the preoccupation of entrepreneurs planning business strategy, which means trying to build a structure and a place in it where it is possible to operate and earn an accounting profit.

Not every investment you mentioned is a sunk cost from the printing entrepreneur’s perspective. The computers he leases are not a sunk cost, because, though they are irrevocably computers, computers are general-purpose devices and can be repurposed to be computers in another business easily enough. Initially installing and setting up the computers in the premises of the printing company would be a sunk cost.

The printing entrepreneur’s commitment to a location, a scale of operation, a particular printing plant, technology and organization, to advertising — all of these involve some sunk cost commitments.

The conventional story about how in the main the printing company expects to recover its sunk cost capital investments revolves around what used to be called, “barriers to entry”. The idea is that subsequent competition will be dissuaded from adding to printing capacity in direct competition and the printing company will enjoy a modicum of something resembling monopoly power, centering on its property ownership of low-cost means of production.

The combination of sunk costs and commitment to fixed costs puts the printing company in the position of a cost structure exhibiting increasing returns to scale, and the company must find enough “market power” to price discriminate. Basically, the printing company cannot just charge its own unit marginal costs, because those are low and declining as output increases, and would never provide enough margin to cover fixed costs, let alone provide a return on sunk cost capital investments. The firm will be motivated to manage its capacity utilization, by selling its printing services at various rates. Advertising and marketing (another mix of fixed and sunk costs, depending upon how you look at it) may be employed. The printing company may join the City Printers Association, where they can cooperate with their competitors to contain destructive competition. There are other clever gambits the printer may try, which I doubt are of much interest.

The larger point I was trying to get at how far afield the neoclassical account of an aggregate production function and a stock of accumulating capital is from the reality of capital as finance for sunk-cost commitments and warranties and insurance. Timing and contingency and power are the essence of capital, and mainstream economist insist on a discourse that puts all that aside.

205

Rakesh Bhandari 12.13.15 at 7:40 pm

Very helpful on connection between r>g and rising inequality. I should not have assumed that people already knew Branko Milanovic’s reply to Debraj Ray.
http://glineq.blogspot.com/2014/06/where-i-disagree-and-agree-with-debraj.html

206

Rakesh Bhandari 12.13.15 at 7:51 pm

One question I have is this: wouldn’t it be possible for capital income to become more concentrated, so that we have a narrow strata of rentiers commanding a greater percentage of capital income and thereby possibly enjoying incomes that dwarf even the bigger ones that come from supermanager/superstar labor…without the capital share of income necessarily rising that much, even if at all? We can get Rastignac’s dilemma becoming more socially relevant without Piketty’s alpha rising sharply, it seems to me.
I am not sure how these two outcomes are related in Piketty’s work–the greater concentration of capital income and the rising share of capital income in total income.

207

Rakesh Bhandari 12.13.15 at 8:16 pm

Here is Debraj Ray’s response to Milanovic
http://debrajray.blogspot.com/2014/06/ray-on-milanovic-on-ray-on-piketty.html
“The second piece of Branko’s reasoning is correct. If we assume that capital incomes are more unequally distributed than labor incomes, and if we assume that the share of capital incomes in the total increases over time, then sure, total inequality will generally climb. But all these “ifs” are endogenous outcomes. Sure, they could be empirically what we observe. And as I have said many times: I fundamentally agree with the empirics. But please, don’t tell me that this is some deep theory of inequality when all the “ifs” in that sentence pretty much assume the answer already.”
So if it is not deep theory of inequality but mostly empirics…

208

Rakesh Bhandari 12.13.15 at 8:36 pm

Yes, I see that people have noted above that Dan Kervick has made similar points to Milanovic. I agree with them.

209

Robert 12.13.15 at 8:38 pm

I was disappointed that Jones’ JEP article did not use the terms “functional distribution” and “personal distribution” of income. I generally think mainstream economists do not draw on ideas arising decades ago.

I have come to believe Piketty draws conclusions not so much from his laws, but from his laws in a concrete setting.

210

notsneaky 12.13.15 at 8:55 pm

Yes, Debraj pretty much nails it.

211

Bruce Wilder 12.13.15 at 9:02 pm

Debraj Ray’s “r > g has nothing to do with any of this” repeated over and over becomes tedious in the extreme. I get that analytic thinkers like to make fine and “precise” distinctions. Sometimes, I like to do that myself. But, if you cannot concatenate your thoughts, you cannot see the world as it is. The world is a product of coincidence, of a combination of factors coming together.

212

dsquared 12.13.15 at 9:12 pm

I have come to believe Piketty draws conclusions not so much from his laws, but from his laws in a concrete setting.

This is exactly right, and if I’d had time to participate in this book event it would have been the general theme of my contribution. Thinking about the r/b and s/beta ratios as long term comparative statics in the normal manner of economics is exactly what I’m currently annoying Radek about. They’re local laws-of-motion, and the passages talking about the long term asymptotic “solutions” aren’t meant to be taken seriously – they’re meant to motivate understanding of those local laws-of-motion and what they might mean for the direction that we’re currently going in. The Jane Austen and Flaubert references seem to be taken by all reviewers with an economics background as ornamental illustrations of general principles, like Krugman’s babysitting circle, but they’re not. They’re meant to be illustrations of particular situations which resemble where we are now. That’s why, despite agreeing with Jamie Galbraith (and, in a different way, with Notsneaky) that the capital theory is for the most part garbage, I think it’s a very important and intellectually significant book.

213

Rakesh Bhandari 12.13.15 at 9:37 pm

Vautrin is a character from Balzac, not Flaubert. By the way, no number of extra tickets can make the parents go out if the poor-quality shows are not worth the price. Right now the shows suck.

214

notsneaky 12.14.15 at 1:26 am

Bruce, he’s repeating it over and over again because people insist on not getting it. And it’s fairly straightforward (which suggests that people, otherwise very smart people, got so invested in Piketty that they’ve forgotten how to think)

Look, this isn’t about some high falutin’ theory. This isn’t about complicated mathematical models. It’s not about abstraction vs. “the real world”. It’s really about basic, capital-L, logic.

You have a proposition. “higher r-g increases inequality”. You need to make an argument in support of this proposition. The conclusion needs to follow from the premises. Some people here are acting like this is some crazy mean thing to demand of Piketty because “empirical work!”. Because “law of capitalism!” Because “data!” Which are all nonsense replies. This is suppose to be a freakin’ scholarly work. The least we can have is an explanation of why the assertion being made is actually true.

But the conclusion doesn’t follow. There is no logic to the conclusion. This is perfectly illustrated in Milanovic’s (whom I otherwise respect very much) faulty explanation, the part which begins with

“returns from capital are privately owned …”

and ending with

“And this happens precisely when r exceeds g”

I’m not quoting the full thing, you can follow Rakesh’s link above and read it yourself. But this is a classic example of a syllogism – the kind they *really* put on Logic 101 exams for freshman – where you have:

True premise. True premise. True premise. True premise. Irrelevant and unwarranted conclusion.

Horses are mammals. You can ride horses. You can ride other mammals. Horses have hooves. Therefore, unicorns haven’t evolved mammary glands yet!!!

The “And this happens precisely when r exceeds g” part is just tacked on at the end and *in no way follows from any of the premises*. Milanovis (and Piketty) just sneak in the conclusion after confusing the reader (and probably themselves) with the bunch of true, but irrelevant premises.

215

notsneaky 12.14.15 at 1:37 am

Daniel, I do think it’s an important and significant book. I haven’t made up my mind if it’s a “good” book. There’s way too much irritating obfuscation and stuff that’s just plain wrong in it (in addition to the r vs. g thing, Piketty also manages to get Malthus, Ricardo and Marx wrong).

It’s not an original sentiment, many people much smarter than me have said it, but if he had just laid off the r vs. g and focused on the data, described the trends historically and provided context-specific explanations (“here wealth inequality decreased because the Bolsheviks refused to pay back Russian debt, here wealth inequality increased because of this housing law that was passed, here inequality increased further but it’s not clear what happened…” etc.) rather than trying (and failing) to uncover some arcane “laws of capitalism” it would’ve been a much intellectually sounder work. Probably wouldn’t have been a best seller though and probably wouldn’t be discussing it at this seminar.

(and let’s be honest, Cit21stC, as scholarly as it is, does have a whiff of those “How to Invest During the Coming Financial Apocalypse” books that show up in your sidebar if you don’t turn your adblocker on about it)

216

Markos Valaris 12.14.15 at 4:39 am

I am not an economist and have not even read Piketty’s book, but being able to follow basic algebra, I am somewhat bemused by this whole debate, including the Ray/Milankovitch exchange.

Ray, and I guess notsneaky here, insist on the point that “income is income is income”, and what matters for how quickly you build wealth is how much of it you save. And Milankovitch et al. insist on the point that how much of your income you save depends on how high your income is, and so if income from profits increases faster than income from labour then those who are already sitting atop a pile of wealth will enjoy faster income growth (and hence higher savings rates) — and so will pull away from the rest.

Can’t all of these things be true together? Indeed aren’t they *actually* all true together?

217

Rakesh Bhandari 12.14.15 at 6:20 am

Milanvoic writes:

“The standard growth requirement that r>g is indeed a contradiction of capitalism because in capitalism, capital owners are private persons and they tend to be rich. And these rich people save more of their income, and gradually build up more and more of their capital stock, thus creating precisely the process of divergence of which Piketty speaks.”

notsneaky counters: “True premise. True premise. True premise. True premise. Irrelevant and unwarranted conclusion.”

Of course this is not Ray’s argument. He grants that Milanovic’s argument is valid. He asserts however that the additional probably true premises–about the concentration of capital, savings behavior and the savings coming from capital income–cannot be derived from the r>g inequality.

It is possible however that Piketty tries to show that the higher r-g, the more shocks are likely to concentrate capital ownership.

But leave that aside, Piketty is clear that given the concentration of wealth, the higher r-g (and it tends to be higher for the wealthy), the more likely it is that after rents are consumed and taxes paid that wealth will grow faster than income. See p. 351, 571

Add to this the competitive reduction of taxes and the slow down of growth from the petering out of catch up growth and demographic slow down and we are well on our way to rentier society.

And of course there are objections. How will labor relations have to be restructured to keep r high enough as more capital is accumulated and what kind of risky financial engineering will be used to create assets with the requisite returns. As Steven Pressman notes, all this make serious social, political and labor crises quite possible along the way to rentier society.

And yes there is a sense of incompleteness about Piketty’s theory because it is not a general theory of how the variables are related. Cooper with whom Kerwick is in dialogue gives a good example. If g falls near zero but r and s remain around 5 and 10%, Piketty’s theory gives us such a high beta that alpha or the capital share of income (which is r times alpha) would be more than the total income. Impossible. Either r or s would have to adjust. But Piketty does not give us a theory of how the adjustments could or would be made given the ways in which the key variables are related.

But still what Kervick is rightly saying is that Piketty gives us a reasonable historical range for these variables to play out, which gives us dynamics over the medium term; and it is the case in fact that nothing is standing in the way of a return to a rentier society.

218

Peter T 12.14.15 at 11:34 am

Rakesh

“If g falls near zero but r and s remain around 5 and 10%,..”

g has been near zero for most of human history. r has always been at least 4 per cent. s is a residual, undefined except by reference to some larger system. Piketty noticed these things. The categories are not universals, but only make sense in terms of particular socio-economic systems. r g and s are not electrons, they are arrangements.

219

dsquared 12.14.15 at 3:07 pm

If g falls near zero but r and s remain around 5 and 10%, Piketty’s theory gives us such a high beta that alpha or the capital share of income (which is r times alpha) would be more than the total income. Impossible. Either r or s would have to adjust.

it’s these adjustments which are the important thing, as Piketty does keep saying. It matters a lot whether the adjustment to r or s comes about as a smooth process of equilibration, or via something like the Russian Revolution. As I read it, the “laws” are really just meant to demonstrate that various kinds of conditions are unsustainable and therefore will, at some point in the future, stop.

220

dsquared 12.14.15 at 3:18 pm

or to put it another way, a lot of the time when he’s talking about relationships between r and g, Piketty is going “look, the system can’t keep going on like this” and everyone’s responding “but it’s impossible for a system to go on like that”.

221

TM 12.14.15 at 4:07 pm

“Piketty is going “look, the system can’t keep going on like this””

Piketty says nothing of the sort. What he says is that the system kept going for 200 years (from 1700 to 1900) with a stable C-I ratio of 700% and he expects capitalism (after that 20th century anomaly) to resume its old rentier mode of operation. This is what makes the debate so fruitless (sigh) – people keep imputing things that simply aren’t in the book.

222

MisterMr 12.14.15 at 4:12 pm

Since the thread has already surpassed the 200 comments, I’ll repeat my opinion.

The problem of Piketty’s argument is that he assumes that it is capital that creates profits, so as long as capital inreases relative to income the profit share also increases, if the profit rate is fixed it will eventually eat up all income.

But this makes no sense because “capital”, in the sense of produced means of production, cannot increase faster than income, because in fact capital produces said income (together with labour of course, but there is a linear relationship between capita and labour too). To put it clearly, capital in the proper sense literally cannot grow faster than income (apart from weird hipotheses about technological change).

However Piketty is not measuring “capital”, but wealth. Since wealth is a fictitious thing, clearly “wealth” can increase faster than total income.

In particular the value of any asset is the capitalized cashflow from that asset at the expected rate of profit, so that if I have a piece of land that gives me 100$/year, and the expected rate of profit is 5%, the value of that piece of land will be 100$/5%= 2000$. From this we can see that, given an expected rate of profit, we can calculate total wealth as:

income X (1- wage share)/(expected rate of profit)

and the wealth to income ratio as:

(1-wage share)/(expected rate of profit)

From this it is obvious that it is not the accumulation of wealth that brings down the wage share, but rather the fall in the wage share that causes an increase of wealth.
It is also obvious that “wealth” is much more than the value of proper capital, hence it is mostly rentier assets and/or financial stuff.
It is also obvious that the r>g thing doesn’t make sense, since the whole thing is driven by the fall in the wage share.

[/rant]

223

Rakesh Bhandari 12.14.15 at 4:13 pm

in reply to @223. Somewhere, perhaps in a talk, Piketty says that he is quite Leninist about the Great War in that he believes that the social tensions of a rentier society found their outlet in the conflagration that changed everything. So I think dsquared is right about the system not being able to go on like this.

224

dsquared 12.14.15 at 4:42 pm

I think we might have seen the same talk, because I remember it from Youtube rather than the book too.

225

notsneaky 12.14.15 at 6:14 pm

“And Milankovitch et al. insist on the point that how much of your income you save depends on how high your income is”

No, that’s what Raj says. Milanovic says that if it’s income from capital you save it, if it’s income from labor you eat it. That’s precisely what Raj critiques with the example of a subsistence farmer who owns a small plot of land (capital) – the fact that they get their income from this small piece of land is not going to make them save any more than if that subsistence income came from wages.

226

notsneaky 12.14.15 at 6:34 pm

@222 – That’s neither what Piketty is saying nor what his critics are saying.

Piketty is unclear whether by “r-g matters” he means that it matters in a “higher level of inequality” kind of way or in a “ever increasing inequality” kind of way. But IIRC he does at at a point or two come out and say it’s NOT the second although he throws out some completely unfounded speculations that “it might”. When he’s actually discussing his “laws” and all that it very much seems like he’s saying “higher level of r-g will lead to a higher level of inequality”

His critics are NOT saying “it’s impossible for the system to go on like that”. They (we?) are saying “r-g has nothing to do with it. The system can have a high r-g and be perfectly stable. The system can have a high r-g and low inequality. r-g is just completely irrelevant”. And it is.

227

Robert 12.14.15 at 6:53 pm

Not sneaky, are you writing r -g, instead of r > g, for any reason other than irritation over HTML trivia?

228

notsneaky 12.14.15 at 7:10 pm

No, it’s irritation over HTML markup

229

dsquared 12.14.15 at 7:47 pm

They (we?) are saying “r-g has nothing to do with it. The system can have a high r-g and be perfectly stable. The system can have a high r-g and low inequality. r-g is just completely irrelevant”

This is where Piketty often gets it wrong because he doesn’t take capital theory seriously, and where his critics encourage him to make the same mistakes as he does. It’s possible to construct all sorts of steady state examples of imaginary economies. But in the real economy living in real historical time, a short term local law of motion is that if r is greater than g, the proportion of output which is being claimed by capital holders has increased over the period. And, also in the conditions we currently live in, capital holders start the period at the top end of the distribution of wealth, and therefore over the period, the amount of output going to people at the top end of the distribution of wealth increases.

The issue then is “how short term” and “how local” this law of motion can be. The answer to the question “what happens in the very long term” depends on assumptions about what capital-owners do with the increment to their share of output, and on elasticities of substitution between factors of production. Piketty often forgets to have his head screwed on and gets into these debates. But actually, the important thing in the book is that “short term” can be “really quite a while, certainly long enough for very significant social pressures to build up”.

And therefore, that the question of the equilibrium properties of any given model is less interesting than you’d think – before anything can be brought back into equilibrium by the operation of marginal substitutability, it’s more likely than not that something like the Great War or the Russian Revolution will have happened, and rebased all the key ratios to a different set of values and a new set of conditions. Collin Street at #204 seems to be making this point – the “economist” critics of Piketty are concerning themselves with the equilibria, but actually the dynamics matter to. The equation mapping wealth distribution and the capital share at time T to their values at time T+1 is a difference equation, and if it’s a difference equation with destabilising positive feedback and damping mechanisms which are weak and/or slow, that’s really important to know.

230

notsneaky 12.15.15 at 1:32 am

” a short term local law of motion is that if r is greater than g, the proportion of output which is being claimed by capital holders has increased over the period”

But there’s absolutely no reason to believe so. I’m not sure what a “local law of motion” is, but quite simply if r is greater than g, the proportion of output being claimed by capital holders can go up or it can go down or whatever. Not just in steady state. “Locally” or “in the short run” or whatever. The assertion above is just that, an assertion without EITHER a theoretical (logical) explanation OR serious empirical support.

Your distinction between “short run” and steady state is a red herring to a red herring. A redherringsquared.

“T to their values at time T+1 is a difference equation, and if it’s a difference equation with destabilising positive feedback and damping mechanisms which are weak and/or slow, that’s really important to know.”

Same question as I asked Collin. What is this difference equation? Both of you are saying “there exists some ephemeral difference equation which if only we knew what it was would prove Piketty right”. Ok, gimme the damn equation and then we can talk.

231

Markos Valaris 12.15.15 at 1:32 am

@227, 228 I feel we are really arguing over rhetorical emphasis here.

If you and I start out with perfectly equal income, but mine comes from wealth and yours comes from labour, and we save the same amount out of our incomes in Period 0, then in Period 1 my income will be higher than yours if and only if r>g. Then I will: (i) save more, because savings rates increase with income; and (ii) have more wealth carried over from Period 0. So, with r-g still the same, our incomes in Period 2 will be even more unequal. And so on, to ever increasing inequality.

Is r>g the whole story? No. Is r>g completely irrelevant to the story? No again.

232

Bruce Wilder 12.15.15 at 1:35 am

notsneaky @ 227: That’s precisely what Raj critiques with the example of a subsistence farmer who owns a small plot of land (capital) – the fact that they get their income from this small piece of land is not going to make them save any more than if that subsistence income came from wages.

And precisely what is Debraj Ray’s point? Does even Ray know? I doubt it.

As soon as he’s gotten past his self-important constipation on this exceedingly fine point, he’s off and running with fuzzy thinking of his own. For example: “If you consistently save a larger fraction of your income than I do, and our labor incomes grow at the same rate g, you will become infinitely richer than me as time goes by. I guarantee it. Yet, at the risk of developing a sore throat, let me say it for the 89th time: r > g has nothing to do with it.”

233

Markos Valaris 12.15.15 at 1:44 am

…@233 scratch that, actually, my example doesn’t work; I think I see the point now.

234

notsneaky 12.15.15 at 1:59 am

Markos, no. If we have the same income and we save the same amount then we’ll both have the same income next period too. It’s just that in next period a part of my income will also be capital income. Again, r-g does not play a role here.

235

notsneaky 12.15.15 at 2:01 am

“And precisely what is Debraj Ray’s point? Does even Ray know? I doubt it.”

Umm, he’s pretty clear about it and even the passage you quote makes it explicit. The point is that r-g is irrelevant to inequality.

How can you both quote it and completely miss it simultaneously?

(Markus, I didn’t see your comment right after mine)

236

Markos Valaris 12.15.15 at 2:40 am

@236 “If we have the same income and we save the same amount then we’ll both have the same income next period too.”

Unless I’m doing the calculation wrong again, it’s actually even worse for my original argument!

If we both saved a in the first period, then my income in the second period will be (i/r+a)*r, while yours will be a*r+(1+g)*i (where i is the income we both enjoyed in the first period).

So for positive g your income will be higher than mine in the second period, regardless of what r is.

237

Rakesh Bhandari 12.15.15 at 5:01 am

Take two individuals whose share of income from capital and labor is the same, but one has much greater total income than the other. Milanovic grants that r>g does not lead to widening inequality in this case, but if the one with larger total income, say 10x greater, had received his income only as capital income, inequality will widen if r>g. Milanovic makes this clear by making the income stream of the richer person entirely capital income. Now his income rises $100 rather than the $80 sum that would have kept the income of the richer person at the same multiple of the poorer one. Inequality increases by .1% in this example.

Check out his example.

This widening of inequality results because r is greater than g. If r was the same as g, say 4%, it would not widen inequality if the richer person were receiving all his income from capital or from some mix of capital ownership and labor. The reason that the switch widens inequality is indeed r>g

Now Debraj Ray responds to this example. But I don’t find his response responsive. Here it is:

“Oh dear, here we go again, a repetition of the very same error I’ve been trying to fix. Let me first make my point in words, and then we’ll see where Branko makes his mistake.

Here is the point: income is income is income, no matter where it comes from. Therefore, inequality will increase if the rich save more than the poor. It will stay constant if the rich save at the same rate as the poor. And it will decline if the rich save at a lower rate than the poor… “

???

238

Bruce Wilder 12.15.15 at 8:05 am

notsneaky @ 237: . . . even the passage you quote makes it explicit. The point is that r-g is irrelevant to inequality.

The passage I quoted repeated the point, which is not quite the same as making it. Repetition of a thesis is rarely proof. And, what Ray put forth as speculative illustration was: “If you consistently save a larger fraction of your income than I do, and our labor incomes grow at the same rate g, you will become infinitely richer than me as time goes by. I guarantee it.” He guarantees it. Wow. But, since we are being so punctiliously precise, let me just point out that if the Saver in Ray’s story is saving into a hoard, for which there is no r, then he won’t become “infinitely richer” as time goes by. He’ll be richer in the event that some Shock to the System puts them both out of work and neither has labor income. But, r being zero, or plausibly close to it, for the investment opportunity of a hoard, the Saver is out of luck as far as the prospect of becoming “infinitely richer” is concerned.

I’m not going to take a brief for Piketty on this. I find the fuzzy, shifting definitions and periods irritating, too. But, Piketty provides plenty of qualification in what he says, and it is possible to understand him on his own terms, without thinking r > g is a theorem to be proved. I think Piketty deserves more respect than Ray is giving him here, and Ray could show more humility.

239

TM 12.15.15 at 9:08 am

“I think Piketty deserves more respect than Ray is giving him here, and Ray could show more humility.”

Piketty deserves a critical reading, and he deserves for his arguments, including the confused and the misguided ones, to be taken seriously, and when necessary to be refuted. I doesn’t seem very respectful to me to say that Piketty is vague and fuzzy but if we read him generously, reinterpret his arguments a bit to our liking and impute some stuff that isn’t actually in the book (e.g. 225 etc.) then he makes a bit of sense, or at least can be made useful to our political purposes. That is more or less what has happened to Marx btw.

240

Markos Valaris 12.15.15 at 10:42 am

Rakesh, the example doesn’t work so simply I think. Here’s my reasoning.

Suppose you have income i from labor in the first period, and I have income 10*i from wealth. (So my wealth must be 10*i/r.) Suppose I save amount a and you save b.

Whether the inequality in wealth grows of not depends on whether our income grow more or less unequal over time.

So in the next period my income is (10*i/r+a)*r and yours b*r+(1+g)*i. The difference between our two incomes, in other words, went from 9*i to 9*i-g*i+(a-b)*r. Whether that’s more or less than the original 9*g depends on whether g*i is greater than (a-b)*r, where a and b are themselves complicated functions of our total incomes. So, no obviously special role for r-g emerges, though it does seem to push in the direction Milanovitch wants.

241

dsquared 12.15.15 at 2:23 pm

I’m not sure what a “local law of motion” is

You are sure because it’s obvious – you just (correctly) don’t want to agree with all the awful mistakes Piketty makes about capital theory. But:

If we define W as the total “wealth” in the economy in the sense Piketty wants, including anything that produces an income, at the beginning of period t=1 (I’m starting at 1 because the font CT uses does zeros that look like os)

Define rt as the compoundable return on this stock over period t.

Define Y as the period t=0 output and gt reasonably normally.

For period 1, the ratio of the part of output claimed by wealth owners to the total is
P1 = (Wr1)/Y

For period 2, the ratio is

P2 = r2 (W(1+r1)) / Y(1+g1)

A bit of arithmetic manipulation and you get the ratio of P2 to P1 to be

(1+r1)r2 / r1 (1+g1)

Which if you then assume that all r’s and g’s are constant gives you

r+ r^2 / r+rg

So if r^2 is greater than rg (ie, r is greater than g), then P2 is greater than P1. But the “steady state” equation is a simplication – and as Piketty’s own evidence shows, not a very accurate one – of the version with all the time subscripts, which is what I’d call the “local law of motion”

But but but – if you define W in this way it isn’t possible to interpret it sensibly as “Capital”, and if you define r in this way you’re ignoring depreciation and the consumption of wealth owners, and the move from rt and gt to steady state is very underjustified. Yes, and Piketty makes all three of these mistakes. But the identity itself is an important one, because it’s telling you something important about compound returns and the importance of wealth in the economy. And equally clearly, the equation describing the evolution of Wr/Y in successive periods is a difference equation, and if you think rt is an increasing function of Wt-1 (which Piketty appears to think, I’m less than convinced myself), it’s going to be one with destabilising feedback.

242

notsneaky 12.15.15 at 5:24 pm

The problem with that is that r*W is part of Y2, hence of g. A simple counter example is when there is only capital income in which case P2/P1=1. It’s trivial, since then capital’s share is always 1, but it does illustrate the basic point that r and g could be anything and shares could be anything, even short term.

Wasn’t it you who mocked people for making “arguments from accounting identities” in a different context? Maybe it was DeLong or Rowe.

243

dsquared 12.15.15 at 7:45 pm

It’s definitely simple, but it’s not a counterexample because in that trivial case, r can’t be greater (or less) than g.

In general, Yg is going to have an r*W part and another part which represents the increase in output which isn’t claimed by capital. And if r is greater than g, then the ratio of that r*W part to the total g is going to be greater than the previous period’s ratio of r*W to Y.

(Note that in Piketty’s definition, W includes things like government bonds, by the way; it certainly isn’t the case that an increment to r*W has to mean an equal increment to Y.)

I’ve never mocked anyone for making arguments from accounting identities, I don’t think. They’re often very useful.

244

Rakesh Bhandari 12.15.15 at 10:47 pm

@242. ” Suppose you have income i from labor in the first period, and I have income 10*i from wealth. (So my wealth must be 10*i/r.) “
The denominator would be 1/r, correct? Is this what you meant to write?

245

Markos Valaris 12.16.15 at 12:04 am

@Rakesh, what I wanted is a term that, when multiplied by r, gives you 10*i. So the numerator is 10*i and the denominator just r.

246

Rakesh Bhandari 12.16.15 at 12:36 am

then r is not the rate of profit–confusing; so what is it?

247

Rakesh Bhandari 12.16.15 at 12:51 am

OK I guess it does not matter whether r as 1.05 or .05, to use Piketty’s number.

I’ll just let r be what you want. Please define a and b, though
“The difference between our two incomes, in other words, went from 9*i to 9*i-g*i+(a-b)*r. Whether that’s more or less than the original 9*g depends on whether g*i is greater than (a-b)*r, where a and b are themselves complicated functions of our total incomes. “

So you are saying that (a-b)r has to be greater than g*i for inequality to widen, but then saying that this does not show importance of r>g for widening inequality, though it points in that direction.

I guess it now depends on what a-b is, so please define a and b.

248

Markos Valaris 12.16.15 at 2:57 am

@Rakesh, r is the rate of return on wealth, no? So if my wealth is w my income from wealth is w*r.

As for a and b, these are supposed to be the amounts (not rates) saved in the first period.

An interesting case is if you let a = b (which assumes that the worker saves the same dollar amount as the capitalist, but let it stand for a moment). Then, regardless of r, inequality *diminishes* in each period, for any g>0.

A bit less unrealistically (but still unrealistically), suppose that the capitalist and the worker save at the same rate s. Then a = 10*b = 10*s*i. Then the condition for widening income spread becomes 9*s*r>g, rather than straight r>g. If you take s to be something like 1/20, then r can be more than twice as high as g and the income spread still diminishes.

So I guess the moral to all this is that r>g may not be totally irrelevant (at least so far as this toy example goes) as notsneaky and others claim, but it does look insufficient to explain the dynamics of inequality.

I should add that I have no strong views about this. It just struck me as odd that people seemed to have so strong and and seemingly contradictory opinions on something that looked like it should be fairly easy to settle.

249

Rakesh Bhandari 12.16.15 at 4:16 am

Or just differentiate between r and r*, the latter being defined as r-c-t.

c is marginal propensity to consume and t is taxes, as Steven Pressman most helpfully writes.

So the condition for widening inequality is r*>g; the net additions to the capital stock will exceed the growth rate. Or in Milanovic’s example we can see that if we switch the richer person’s income to capital income alone, inequality will widen due to r*>g.

At any rate, you are now disagreeing with S neaky N ot, and stand accused of making logical mistakes a freshman and an even more “rigorous” tenurable economist should not make, even if Piketty and Milanovic got away with it.

A lot of the drama in Piketty’s book is about why g will likely revert to a historical low level and why the competitive inter-state system will reduce t, which is part of the reason we are likely to get r*>g, which is a powerful tendency for growing inequality.

250

dsquared 12.16.15 at 8:06 am

then r is not the rate of profit–confusing; so what is it?

An annoying thing about the book is that this is only the second most confusing piece of not-quite-standard terminology in it, after “Capital”.

251

Markos Valaris 12.16.15 at 9:07 am

Good thing I’m not an economist then, even a freshman one at that!

I don’t really follow your comment, though; what are the logical mistakes? And, I’m not sure if I agree or disagree with notsneaky, Ray, or Piketty/Milanovitch; I’ve only worked through one simple example in which r>g seems to be *one* condition for widening inequality but far from a sufficient one.

252

Rakesh Bhandari 12.16.15 at 7:53 pm

Sneaky Not has been insistent that anyone who insists on r>g ever having anything to do with widening inequality is illogical and foolish. So you have that to live with now.
Of course Piketty does not think r*>g is helpful for understanding the total dynamics of inequality; he does not think it throws much light on the inequality in labor income, for example.
Still I am wondering does the use of my r* get at what I think was part of your concern, viz. that r being greater than g need not widen inequality if the capitalists save very little and become spendthrifts? Or as you write: “If you take s to be something like 1/20, then r can be more than twice as high as g and the income spread still diminishes.”

Comments on this entry are closed.